applications of machine to machine communication in remote
TRANSCRIPT
Applications of Machine to Machine Communication in Remote Healthcare Systems
SAJAD ABED ALMAHDI ABEDALI ALABADI
A thesis submitted in partial fulfilment of the requirements for the degree of Doctor of Philosophy
from University of Greenwich
December 2017
i
DECLARATION
I certify that the work contained in this thesis, or any part of it, has not been accepted in
substance for any previous degree awarded to me, and is not concurrently being submitted for
any degree other than that of Doctor of Philosophy being studied at the University of
Greenwich. I also declare that this work is the result of my own investigations, except where
otherwise identified by references and that the contents are not the outcome of any form of
research misconduct.
Signed:
Student ……………………………………
First Supervisor ……………………………
ii
ACKNOWLEDGEMENTS
I would like to dedicate my dissertation work to my supervisors and family for their
unconditional support. I would like to express my deep gratitude to my first supervisor, Dr.
Ruiheng Wu, for his patient guidance and encouragement, and for the advice, he has provided.
I have been extremely lucky to have a supervisor who cared so much about my work, and who
always impressed me with his wealth of knowledge and his high ethical, academic and personal
standards; without him, none of my success would be possible.
I am heartily thankful to my second supervisor, Dr. Yehdego Habtay, for helping me to develop
my research skills and effective means of communication, and for his independent and
encouraging comments on my work. Dr. Yehdego Habtay’s feedback and guidance have been
incredibly empowering for me, on this important journey of my life. Also, I would like to thank
Professor Predrag Rapajic and Dr. Kamran Arshad for their support at the beginning of my
research journey. Finally, a special feeling of gratitude to my loving parents and brothers,
Shatha and Abed whose words of encouragement and push for tenacity ring in my ears, I thank
them for their endless love.
iii
ABSTRACT
Wireless Machine-To-Machine (M2M) communications in healthcare systems will play a large
part in the medical sector in the near future, enabling a number of applications to speed up
medical treatment, decreasing the cost, and increasing the flexibility and efficiency of the
hospitals and medical bodies. The work undertaken in this thesis is to improve the spectrum
and energy efficiency of M2M communications in the medical sector by exploiting Cognitive
radio (CR) technology.
First, The thesis considered an efficient aggregation-based spectrum assignment algorithm for
Cognitive Machine-To-Machine (CM2M) networks. The proposed algorithm takes practical
thresholds including Co-Channel Interference (CCI) among CM2M devices, interference to the
Primary Users (PUs), and Maximum Aggregation Span (MAS) into consideration. Simulation
results clearly show that the developed algorithm outperforms State Of The Art (SOTA)
algorithms in terms of network capacity and spectrum utilization. The developed algorithm can
improve data rate of CM2M devices by at least 23% compared with the SOTA algorithms.
Furthermore, this thesis presents an optimal energy efficient spectrum management mechanism
with multiple thresholds. The developed mechanism aims to reduce energy consumption in the
system by optimizing spectrum sensing and channel switching, while at the same time
decreasing the probability of collision and assuring the reliability thresholds, throughput, and
delay. Subsequently, an Antenna Selection Sensing (ASS) scheme is used to improve sensing
accuracy. The simulation results show that the energy efficiency of the CM2M gateways can
be improved by at least 35%.
In addition, the thesis considered an Energy-Efficient Channel Selecting (EECS) algorithm for
CM2M communications in the healthcare system. The proposed algorithm aims to select the
best available channels to improve CM2M communication quality and reduce energy
consumption in the system overall. Accordingly, the algorithm reduced the probability of
CM2M gateways switching between available channels and improved the energy efficiency by
at least 45%. The efficiency of the algorithm is discussed and demonstrated through
simulations.
iv
CONTENTS
DECLARATION i
ACKNOWLEDGEMENTS ii
ABSTRACT iii
CONTENTS iv
FIGURES viii
TABLES x
ABBREVIATIONS xi
SYMBOLS xiv
Chapter 1 Introduction
1.1 Background 1-2
1.2 M2M Communications in Healthcare Applications 1-3
1.3 Cognitive Radio 1-4
1.4 Cognitive Machine-to-Machine 1-5
1.5 CR Motivations in M2M E-Healthcare Application 1-6
1.5.1 Number of Healthcare M2M Devices 1-6
1.5.2 Battery Life and Green Technologies 1-6
1.5.3 Support Remote Area Devices 1-6
1.5.4 Avoidance and Interference Reduction 1-7
1.5.5 Data Rate 1-7
1.6 Aim and Objectives 1-8
1.7 Thesis Contributions 1-8
v
1.8 List of Published Works 1-9
1.9 Thesis Outline 1-10
References 1-12
Chapter 2 Cognitive Machine-to-Machine Background
2.1 Introduction 2-2
2.2 Machine to Machine in Healthcare Applications 2-2
2.3 Cognitive Radio Spectrum Sensing and Sharing 2-7
2.3.1 Spectrum Sensing 2-7
2.3.2 Transmission Detection 2-8
2.3.3 Cooperative detection 2-9
2.3.4 Spectrum sharing
2-10
2.4 Spectrum Aggregation 2-10
2.5 Spectrum Sensing and Sharing 2-12
2.6 Learning and Channel Selecting 2-13
2.7 Summary 2-14
References 2-16
Chapter 3 Aggregation Based Spectrum Assignment for Cognitive
Machine-to-Machine (M2M) Network
3.1 Introduction 3-2
3.2 Spectrum efficiency (Systems Model) 3-4
3.2.1 Spectrum Assignment Model 3-4
3.2.2 Spectrum Aggregation Model 3-7
3.3 Problem Formulation 3-10
vi
3.3.1 Optimisation problem 3-10
3.3.2 Spectrum Aggregation Algorithm Based on Genetic Algorithm 3-11
3.4 Simulation Results 3-14
3.4.1 Scenario-I: Without Co-Channel Interference 3-16
3.4.2 Scenario-II: With Co-Channel Interference 3-17
3.4.3 Convergence of MSRA 3-19
3.5 Summary 3-21
References 3-23
Chapter 4 Energy Efficient Spectrum Sensing and Sharing
4.1 Introduction 4-2
4.2 Spectrum Sensing and Sharing System Model 4-3
4.3 System Problem and Solution 4-8
4.4 Simulation and Discussion 4-11
4.5 Summary 4-13
References 4-14
Chapter 5 Selective Antenna Sensing
5.1 Introduction 5-2
5.2 Selective Antenna Sensing System Model 5-3
5.3 Selective Antenna Sensing System Problem 5-5
5.4 Problem Solution 5-6
5.5 System Simulation 5-7
5.6 Summary 5-11
References 5-12
vii
Chapter 6 Energy Efficient Scheduling Algorithm
6.1 Introduction 6-2
6.2 Energy Efficient Scheduling Algorithm (System Model) 6-4
6.3 Simulation and Results 6-8
6.4 Summary 6-2
References 6-3
Chapter 7 The Future of M2M in The Healthcare Sector
7.1 Introduction 7-2
7.2 The Factors Driving M2M Adoption 7-3
7.3 M2M Opportunities in the Healthcare Sector 7-3
7.4 Real World Examples 7-6
7.5 M2M Applications: Future Challenges 7-10
7.6 Security Risk and Vulnerabilities in Healthcare Sector 7-11
7.7 Summary 7-14
References
7-15
Chapter 8 Conclusion and Future Work
8.1 Conclusion 8-2
8.2 Future Work 8-3
References
8-5
Appendix A
Appendix B
viii
FIGURES
1.1 Machine-To-Machine Devices Usage Expectations 1-2
1.2 M2M Typical Network 1-3
1.3 Remote Area Communications of Typical M2M Network 1-6
1.4 Flowchart of Thesis 1-10
2.1 ETSI System Architecture for M2M Networks in Healthcare Systems 2-3
2.2 CM2M Activities with Channel Transmitting 2-7
3.1 Subcarrier Distribution Over Spectrum 3-3
3.2 Architecture Diagram of CM2M Network Operating in TVWS 3-5
3.3 Aggregation of Disjointed Spectrum Fragments 3-7
3.4 MSR Algorithm Flow Chart 3-12
3.5 The Impact of Varying Network Load Conditions on Spectrum Utilisation (Scenario-I: Without CCI)
3-17
3.6 The Impact of Varying Network Load Conditions on Spectrum Utilisation (Scenario-II: With CCI)
3-18
3.7 The Impact of Varying Network Load Conditions on Number of Rejected CM2M Devices (Scenario-II: With CCI)
3-19
3.8 The Impact of Number of Generations on The MSRA Results 3-20
3.9 Distribution of Processing Time for MSRA to find an Optimal Solution
3-21
4.1 CM2M System with a Number of Secondary Transmission CM2M Gateways
4-4
4.2 CM2M Gateways Performance in N Packet of Data 4-5
5.1 CM2M Network in Healthcare System 5-2
5.2 CM2M Gateways Performance in Transmitting Ɲ Packet of Data
5-4
5.3 Using Antenna Selection Sensing (J = 2) 5-9
ix
5.4 𝐋𝐬 Bounds of CM2M gateways as a function of 𝛕𝐬 5-10
5.5 The Energy Consumption of The CM2M Gateways Under Various Sensing Mechanisms
5-11
6.1 Energy Efficient Scheduling Algorithm Flowchart 6-3
6.2 The Outcomes of Channel Selecting Probability Based on EECS Algorithm
6-10
6.3 Staying Probability at 𝛕𝐬 Optimal 6-11
6.4 Energy Consumption at 𝛕𝐬 Optimal 6-12
7.1 Telehealth Usage 7-2
7.2 HeartAssist5, Numerex 7-6
7.3 SmartSole, GTX Corp device 7-8
x
TABLES
3.1 Simulation Parameters 3-15
3.2 System Parameters 3-20
4.1 Value Setting of Simulations 4-11
4.2 Energy Consumption in CM2M Devices 4-12
4.3 Energy cost when optimal 𝑳𝒔 and 𝝉𝒔 used, compared with the one in
which employed 𝝉𝒔 only
4-13
5.1 Value Setting of Simulations 5-8
6.1 Value Settings of Simulations 6-9
7.1 CM2M Communications Security Requirements in Medical Applications Sector
7-13
xi
ABBREVIATIONS
xii
(3GPP) 3rd Generation Partnership Project
(AASAA) Aggregation Aware Spectrum Assignment Algorithm
(ABI) Allied Business Intelligence
(AI) Artificial Intelligence
(ASS) Antenna Selection Sensing
(CA) Carrier Aggregation
(CCI) Co-Channel Interference
(CM2M) Cognitive Machine to Machine
(CR) Cognitive Radio
(CRN) Cognitive Radio Network
(CSCG) Circularly Symmetric Complex Gaussian
(COPD) Chronic Obstructive Pulmonary Disease
(CHF) Congestive Heart Failure
(DOFDM) Discontinuous Orthogonal Frequency Division Multiplexing
(DSA) Dynamic Spectrum Access
(EWA) Experience-Weighted Attraction
(EIU) Economist Intelligence Unit
(ECG) Electrocardiogram
(EE) Energy Efficient
(EECS) Energy Efficient Channel Selection
(EMI) Electromagnetic Interferences
(ETSI) European Telecommunications Standards Institute
(EWAH) EWA Handoff (EWAH)
(FCC) Federal Communications Commission
(FDA) Food and Drug Administration
(FQL) Fuzzy Q Learning
(GA) Genetic Algorithm
(IBSG) Internet Business Solutions
(IoT) Internet of Things
(ISM) Industrial, Scientific, And Medical
(LTE-A) Long-Term Evolution Advance
(LUs) Licensed Users
(M2M) Machine to Machine
(MAS) Maximum Aggregation Span
xiii
(MSA) Maximum Aggregation Span
(MSR) Maximising Sum of Reward
(MBAN) Medical Body Area Network
(MSRA) MSR Algorithm
(OFDM) Orthogonal Frequency-Division Multiplexing
(OFDMA) Orthogonal Frequency-Division Multiplexing Access
(PN) Primary Network
(PUs) Primary Users
(QoS) Quality of Service
(QPSK) Quadrature Phase Shift Keying
(RCAA) Random Channel Assignment Algorithm
(SA) Spectrum aggregation
(SAS) Single Antenna Sensing
(SDR) Software Defined Radio
(SNR) Signal-To-Noise Ratio
(SOTA) State-Of-The-Art
(SUs) Secondary Users
(TBS) Turn Based Strategy
(TVWS) TV White Spaces
(UHF) Ultra-High Frequency
(WMTS) Wireless Medical Telemetry Service
(WRAN) Wireless Regional Area Network
(WSN) Wireless Sensor Networks
xiv
SYMBOLS
xv
B Reward vector
𝑳𝑳 Channel availability matrix
ɸN 𝑛𝑡ℎ CM2M devices number
y𝑀 𝑚𝑡ℎ Channel number
N CM2M devices
M Non-Overlapping orthogonal channels
ɼn Available channels at ɸN location
𝑨𝑨 The interference constraint matrix
𝑅 Device requested bandwidth vector
𝑟𝑛 Bandwidth demand of ɸN
Δ𝑓 Bandwidth of sub-channel
𝑤𝑛 The number of requested sub-channels by ɸn
₦ The set of natural numbers
𝐵𝑊𝑚 The bandwidth of 𝑦𝑚
ℱ𝑖,𝑚,𝐿 Lowest frequency of ��𝑖, 𝑚
ℱ𝑖,𝑚,𝐻 Highest frequency of ��𝑖, 𝑚
𝑎 Index of each sub-channel within the available spectrum
DD The sub-channel assignment matrix
|. | Cardinality of a set
U Network utilisation function
𝑘𝑚 Number of sub-channels in channel in 𝑦𝑚
BW Channel bandwidth
��𝑖, 𝑚 𝑖𝑡ℎ Sub-Channel of 𝑦𝑚
Ƚ Network load
𝑃𝐿 Path loss model
𝑢 The ratio of the sum of rewarded bandwidth to the sum of all available
bandwidth
T Transmission slot duration
𝜏𝑠 Sensing slot duration
𝑃𝑓 False alarm probability
𝑃𝑑 Detection probability
𝑓𝑠 Sampling frequency
𝑓𝑐 Carrier frequency
xvi
��𝑓 Target false alarm probability
��𝑑 Target probability of detection
𝜏𝑠𝑚𝑖𝑛 Minimum sensing duration to achieve the reliability probabilities constraints
(1 – 𝐿𝑠) The probability that CM2M gateways will switch to another free channel
𝐿𝑠 The probability that CM2M gateways will wait and sleep in the current
sensed channels
X Overall energy cost needs to finish transmitting one packet of data
𝑆 The time needs for the CM2M gateways to send a packet of data
ℛ The average data rate or throughput
𝑟 The minimum throughput that should be achieved in the system
𝑃3 The switching probability to another idle channel
𝑃1 The probability of a channel being idle
𝑃𝑏 The probability of a channel being busy
𝑝𝑓𝑐 Probability of collision between CM2M gateways
𝐽𝑠𝑤 The energy cost for single channel
𝐸𝑡 Power cost per second due to sending data
𝐸𝑠 Power cost per second due to sensing process
𝐶0 M2M gateways throughput
𝑃𝑒 The probability of the CM2M gateways switching to a busy channel
𝐵𝑡 The number of bits transmitted in one transmission slot
𝑃𝑐 The probability of channels correctly sensed as busy
𝐿𝑠𝑜𝑝𝑡
Optimal sleep probability
𝑅0 The average throughput of CM2M gateways
Q CM2M gateways number
𝜇 Throughput coefficient
ᶆ Random channels shared between secondary users and primary users
Ɲ The number of frames
∆ (𝜏𝑠) Feasibility region
J Number of antenna
Z RF chains
𝑃𝑥 The probability that the CM2M gateways will choose to switch to an ideal
channel when the current channel is occupied, and at least one of the other
channels is free.
xvii
𝛼К The idle probability of channel к
𝛽К The successful transmission probability of channel к
y The probability of selecting channel
γ SNR regime
µ Attenuation coefficients
𝜎 Attenuation coefficients of probability
Γ The channel available probability vector
𝐼[⋅] The indicator function
Page 1 - 1
CHAPTER 1
INTRODUCTION
Page 1 - 2
1.1 Background
Machine-to-machine communication can be defined as communication which can intelligently
process and transfer data independently, without human intervention [1]. M2M communication
enables the exchange of information between networked devices and business application
servers and is considered an important part of the Internet of Things (IoT).
The number of machines connected to the internet has increased dramatically in recent years.
In 2012 there were 8.7 billion machines/devices connected to the internet, and due in part to
the growth of smart devices and tablet PCs, the number of devices reached 28.4 billion in 2017
[2]. Cisco Internet Business Solutions Group (IBSG) predicts M2M devices will increase
massively in next few years, there will be 42.1 billion devices by 2019 and 50.1 billion by
2020, as shown in Figure 1.1 [2].
These devices are expected to be massively used in a number of applications, including
agricultural and industrial automation, healthcare, automobiles, metering and control of
electricity, gas, heat, and water, etc. [3].
Figure 1.1: Machine-To-Machine (M2M) Devices Usage Expectations
0
10
20
30
40
50
60
2015 2017 2018 2020
M2M Devices in Billions
Page 1 - 3
M2M network architecture usually consists of a combination of wireless networks (Figure 1.2)
that collaborate in collecting, processing and analyzing information [4]. The important parts of
the M2M networks are illustrated as follows:
1. M2M Device Domain: installed at different places which send/receive data based on
actions that occur due to changes in parameters of sensors connected to an M2M
network.
2. Network Domain: represents different applications that receive, analyze, process and
understand the data received from the sensors.
3. Application Domain: usually consists of IT applications, programs, and billing
systems.
Figure 1.2: M2M Typical Network [6]
1.2 M2M Communications in Healthcare Applications
M2M in e-healthcare can help the current monitoring and healthcare services, particularly for
children, the elderly, and the disabled people. A huge number of advantages can be achieved
by using M2M communications in the healthcare system, for example, by improving the
monitoring capability of hospitals and e-healthcare centers, and by, sending emergency
information to the patient’s doctors or carers.
Page 1 - 4
Healthcare applications in most cases can be defined as a soft real-time system, in which some
latency is allowed [4, 5]. Coordination in emergency conditions, such as with sudden falls and
heart attacks, is vital when it comes to saving lives. Therefore efficient and reliable
communication between e-healthcare systems is required.
Recently, a huge amount of effort has been focused on wireless communication technologies
for e-healthcare applications, such as electrocardiogram (ECG), blood pressure, oxygen, and
temperature monitors. M2M wireless communication (Figure 1.2) is the foremost technology
for enhancing service flexibility and mobility for different e-health applications [4], which
leads to improved flexibility and efficiency in the e-healthcare systems, allowing patients an
earlier discharge from hospitals and a faster return to their normal lives, while reducing the
total cost for both the patient’s family and government spending [6].
This can be done by enabling hospital employees to remotely monitor and access patient data,
increasing the accuracy of diagnoses by collecting data from different sources (e.g., therapeutic
devices, images, and monitors) over time to build a clear and complete picture of the patient’s
health status.
1.3 Cognitive Radio
Cognitive radio is a radio which is aware of its operational and geographical environment,
established policies, and its internal state. It can dynamically and autonomously adapt its
operational parameters and protocols and to learn from its previous experience [7].
Cognitive radio (CR) was developed to utilize the best available wireless channels in its
vicinity, which is very important technology for the realization of M2M communication in
future due to the limited availability of spectrum, especially with the explosion of the Internet
of Things [8]. The idea of cognitive radio is to share the spectrum between licensed users,
called primary users (PUs), and unlicensed users, called secondary users (SUs). CR is based
on Dynamic Spectrum Access (DSA), a new spectrum sharing method that allows SUs to
accessible spectrum portions within the licensed spectrum bands.
CR technology can solve spectrum limitation problems and can be utilized to increase energy
efficiency in M2M networks and prolong the battery life of sensors. [7]. TV broadcast networks
and cellular networks could be classified as PN, while CR networks can be classified as either
infrastructure-less or infrastructure-based. The infrastructure-less cognitive radio CR Ad-Hoc
Page 1 - 5
networks can be used without infrastructure support (no central entity required), while the
infrastructure-based networks have a master network entity, such as a base station or access
point.
1.4 Cognitive Machine-to-Machine (CM2M)
Cognitive radio as a smart technology can address spectrum limitation and energy consumption
problems in M2M networks. CR has recently become a zone of the most significant subjects
in wireless communications [3].
A number of M2M applications can benefit from the new functionality and performance that
can be achieved by the combination of cognitive radio and M2M communications, ranging
from smart grid and healthcare to car parking [11]. CM2M networks can improve spectrum
utilization and energy efficiency in M2M devices [12]. M2M devices in CM2M networks will
act as SUs which can interact with the radio environment by performing spectrum sensing and
accessing available channels. Normally M2M devices (e.g., Medical application sensors) will
analyze the features of the free channels, following which spectrum assignment is made
according to the needs of the M2M devices [13].
1.5 CR Motivations in M2M E-Healthcare Applications
1.5.1 Number of Healthcare M2M Devices
M2M communication will face challenges due to the vast number of M2M devices in the e-
healthcare sector, such as wearable monitors and other medical sensors. The challenges relating
to the number of connected devices need to be resolved using smart technologies. CR as smart
technology can be programmed to deal with a huge number of devices and to manage the
communications between the sensors domain and the applications domain [14, 15].
1.5.2 Battery Life and Green Technologies
Green means: less energy usage, and less CO2. Having a huge number of machines connected
and interconnected wirelessly all the time is affecting the ecosystem. Hence, the standards
bodies are facing problems on two fronts: electromagnetic pollution and excessive power
consumption [16]. M2M devices (especially body sensors) need to be designed to run for a
very long time without any battery replacement. Therefore, energy efficient schemes for those
Page 1 - 6
devices is necessary for the enhanced performance of the M2M communication systems
[12][17]. CR radio proved to be energy efficient and able to decrease electromagnetic pollution
and energy consumption in its network [17][18]. Such a technology can be utilized to tackle
the energy problems in M2M communications.
1.5.3 Support Remote Area Devices
Cognitive M2M networks can exploit wireless technology to operate in remote areas and solve
the problem of many companies that still grapple to gain a good service (e.g., with internet
provision). CM2M networks as smart technology could access available channels and easily
be configured. Therefore, CM2M could be used to reach remote areas as shown in figure 1.3
by exploiting lower frequencies (470-790 MHz) in the Ultra High Frequency (UHF) band,
called TV White Spaces (TVWS) [19]. A number of M2M medical applications today could
work in the TV white space spectrum using CR technology as TVWS can provide simple
connectivity to remote areas with less expensive communications infrastructure [20, 21].
Figure 1.3 Remote Area Communications of Typical M2M Network [21]
1.5.4 Avoidance and Interference Reduction
Efficient techniques are needed to address the problem of interference as M2M devices are
expected to number more than 50 billion by 2020, the majority of which will need to
communicate wirelessly. Smart CR radio can be adjusted to cope with interference problems
using spectrum broker technology, for example by avoiding interference with the primary users
Page 1 - 7
or with the other CM2M devices [22]. CM2M networks based on the IEEE 802.22 standard
offer the chance to automatically find and utilize unused spectrum so CM2M users will be able
to switch between various wireless channels.
1.5.5 Data Rate
Cognitive radio as smart technology can utilize idle spectrum to improve spectrum efficiency,
which can significantly improve data rate and throughput [22][11]. CR self-configurations
enable devices to learn from the past, for example, a cognitive radio can be adjusted to select
the best quality of services available channels after testing the channels for a specific period,
enabling M2M devices to send more data and achieve better throughput.
1.6 Aim and Objectives
This thesis aims to study M2M communications and to propose and develop efficient
mechanisms to improve spectrum utilization and to reduce the energy consumption of M2M
communications in e-healthcare applications. The main objectives of the thesis are listed
below: -
1. Spectrum efficiency: Develop and design new algorithms to improve spectrum efficiency
in e-healthcare M2M communications by exploiting CR technology. The developed
algorithms will consider aggregation-based spectrum assignment algorithms for better
spectrum efficiency and better channel utilization. Furthermore, genetic algorithms will
take into consideration to maximize channels utilization.
2. Energy efficiency: Develop and design new schemes for better Energy Efficiency (EE) in
e-healthcare M2M communications by considering CR spectrum accessing and sharing
algorithms. The developed algorithm will consider sleep modes and switch modes schemes
for better energy and spectrum efficiency. Furthermore, antenna selection algorithms and
learning algorithms will be exploited to maximize the energy efficiency in e-healthcare
M2M communications.
Page 1 - 8
1.7 Thesis Contributions
The contributions of this thesis cover various aspects of M2M communications in healthcare
systems. The key outcomes of this research in the form of novel solutions, algorithms, and
mechanisms are summarised below.
1. An aggregation spectrum assignment for CM2M gateways is designed as a mixed integer
optimization problem. The maximum aggregation span and the practical constraints of Co-
Channel Interference in aggregation aware CM2M networks are taken into consideration.
2. A fast convergence, simple and robust algorithm, called a genetic algorithm (GA), is used
to resolve the aggregation aware spectrum assignment and to produce better results. GA
algorithm guarantees the results improve by at least 23 % compared with the current state
of art algorithms.
3. An optimal energy efficient spectrum management mechanism formulates the number of
constraints on the throughput, delay, and reliability of CM2M devices sensing. The
proposed mechanism aims to decrease energy consumption in the system (by at least 35%)
using efficient spectrum sensing and channel switching techniques, while reducing the
probability of collision, with the guarantee of conforming to throughput, delay and
reliability constraints. Subsequently, an antenna-selection sensing mechanism is used to
improve sensing accuracy and reduce the probability of collision. The optimality of the
used mechanisms is demonstrated through Matlab simulations.
4. An energy efficient channel selection algorithm has been formulated to work with CM2M
communications in the e-healthcare system. The proposed algorithm aims to select the best
available channels and reduce energy consumption in the system. Furthermore, the
algorithm aims to reduce the probability of CM2M gateways switching the available
channels. The efficiency of the used algorithm is shown through simulations.
Page 1 - 9
1.8 List of Published Works
[1] S. Alabadi, Predrag Rapajic, and Kamran Arshad (2016) ‘’Energy-Efficient Cognitive
M2M Communications’’, International Journal of Interdisciplinary
Telecommunications and Networking, vol. 8, no. 3, pp. 1-9, 2016. [Journal]
[2] S. Alabadi, Predrag Rapajic, and Kamran Arshad (2016) “Efficient Cognitive M2M
Communications”, Wireless Telecommunications Symposium, April 18-20, London,
UK. [Conference]
[3] S. Alabadi, S. Rostami, Kamran Arshad and Predrag B. Rapajic (2016) Spectrum
Assignment Algorithm for Cognitive Machine-to-Machine Networks, Hindawi Special
Issue Smart Spectrum Technologies for Mobile Information Systems. [Journal]
[4] S. Rostami, Sajad Alabadi, Kamran Arshad, and Predrag Rapajic (2016) Efficient Sub-
Carrier Allocation Algorithm for OFDM based Wireless Systems, 2016 Universal
Technology Management Conference, Bemidji State University, Minnesota, USA,
May 26-28- 2016. [Conference]
[5] S. Alabadi, Ruiheng Wu and Yehdego Habtay (2017) “Energy Efficient CM2M
Communications in Healthcare Systems”, International Conference on
Telecommunications and Signal Processing (TSP), IEEE, Spain. [Conference]
[6] S. Alabadi, Ruiheng Wu and Yehdego Habtay (2017) “CM2M Energy Efficient
Channel Selecting Algorithm for Medical Applications”, International Conference for
Internet Technology and Secured Transactions (ICITST-2017), IEEE, University of
Cambridge. [Conference]
1.9 Thesis Outline
The rest of the thesis is organized as follows: In Chapter 2 the literature review is presented
and discussed; this chapter also presents related work on different aspects of Cognitive M2M
communications in e-healthcare systems. The main contributions of the thesis, relating to three
distinct areas, M2M gateways spectrum sensing, sharing, and accessing; spectrum
aggregations; and selecting algorithms are discussed in Chapters 3, 4, 5 and six respectively.
Each chapter addresses a unique research problem, as illustrated in Figure 1.4. Based on the
overall picture of research conducted in the thesis, the future of M2M communications in the
healthcare system and, the main conclusions together with some directions for future work are
presented in Chapter 7 and 8.
Page 1 - 10
Figure 1.4: Flowchart of Thesis
Chapter 2
Literature Review
Energy Efficiency
Chapter 4
Energy Efficient Spectrum Sensing and Sharing
Chapter 5
Selective Antenna Mechanism
Chapter 6
Energy Efficient Channel Selection
Spectrum Efficiency
Chapter 3
Aggregation Based Spectrum
Assignment
Chapter 7
The Future of M2M Communications in
Medical Sector
Chapter 8
Conclusion and Future Work
Page 1 - 11
References for Chapter 1
[1] M. Hatton, The Global M2M Market in 2013. London, U.K.: Machina Research
White Paper, Jan. 2013.
[2] Internet of Things (IoT) connected devices installed base worldwide from 2015 to
2025(in billions):https://www.statista.com/statistics/471264/iot-number-of-connec-
ted-devices world wide/. [Online; accessed 07-May-2013].
[3] Aijaz, A., & Aghvami, A.H. (2015). Cognitive Machine-to-Machine
Communications for Internet-of-Things: A Protocol Stack Perspective. Internet of
Things Journal, 2(2), 103-112.
[4] ETSI, Machine to Machine Communications (M2M): Use cases of M2M
applications for eHealth, ETSI TR 102 732, 2011.
[5] X. Li, R. Lu, X. Liang, X. Shen, J. Chen, and X. Lin, \Smart Community: An Internet
of Things Application," IEEE Commun. Mag., vol. 49, no. 11, pp. 68-75, 2011.
[6] ETSI M2M, E-health architecture: Analysis of user service models, technologies,
and applications supporting eHealth, TR 102 764, 2009.
[7] Van-Tam Nguyen,1 Frederic Villain,2 and Yann Le Guillou, “Cognitive Radio RF:
Overview and Challenges,” VLSI Design, Volume 2012 (2012), Article ID 716476,
13 pages.
[8] M. Nekovee, Quantifying the Availability of TV White Spaces for Cognitive Radio
Operation in the UK," in IEEE International Conference on Communications (ICC)
Workshops, 2009, pp. 1-5.
[9] J. Unnikrishnan and V. V. Veeravalli, “Algorithms for dynamic spectrum access
with learning for cognitive radio,” IEEE Trans. Signal Process., vol. 58, no. 2, pp.
750 –760, Feb. 2010.
[10] L. Giupponi, A. Galindo-Serrano, P. Blasco, and M. Dohler, “Positive networks: an
emerging paradigm for dynamic spectrum management [dynamic spectrum
management],” IEEE Wireless Commun., vol. 17, no. 4, pp. 47 –54, Aug. 2010.
[11] Starsinic, M.; "System architecture challenges in the home M2M network,"
Applications and Technology Conference (LISAT), 2010 Long Island Systems, vol.,
no., pp.1-7, 7-7 May 2010.
Page 1 - 12
[12] F. J. Lin, Y. Ren and E. Cerritos, "A Feasibility Study on Developing IoT/M2M
Applications over ETSI M2M Architecture," 2013 International Conference on
Parallel and Distributed Systems, Seoul, 2013, pp. 558-563.
[13] A. Bicen, O. Akan, and V. Gungor, \Spectrum-Aware and Cognitive Sensor
Networks for Smart Grid Applications," IEEE Commun. Mag., vol. 50, no. 5, pp.
158-165, 2012.
[14] W. Zhao, W. Chaowei, and Y. Nakahira, ``Medical application on the Internet of
Things,'' in Proc. IET Int. Conf. Commun. Technol. Appl. (ICCTA), Oct. 2011, pp.
660-665.
[15] Furong Huang, Wei Wang, Haiyan Luo, Guanding Yu, and Zhaoyang Zhang.
Prediction-based spectrum aggregation with hardware limitation in cognitive radio
networks. In Vehicular Technology Conference (VTC 2010-Spring), 2010 IEEE
71st, pages 15, 2010.
[16] N. Accettura, M. Palattella, M. Dohler, L. Grieco, and G. Boggia, "Standardized
power-efficient & internet-enabled communication stack for capillary M2M
networks," in Wireless Communications and Networking Conference Workshops
(WCNCW) 2012 IEEE, 2012, pp. 226-231.
[17] Quoc Duy Vo, Joo-Pyoung Choi “Green Perspective Cognitive Radio-based M2M
Communications for Smart Meters”, Tutorials, IEEE 978-1-4244-98072010.
[18] Lu, R., Li, X., Liang, X., & Lin, X. (2011). GRS: The green, reliability, and security
of emerging machine to machine communications. IEEE Communications
Magazine, 49(4), 28–35. doi:10.1109/MCOM.2011.5741143.
[19] Pero Latkoski, Jovan Karamacoski, and Liljana Gavrilovska. Availability
Assessment of TVWS for Wi-Fi-like Secondary System: A Case Study. In Cognitive
Radio Oriented Wireless Networks and Communications (CROWNCOM 2012),
pages 196–201, Stockholm, June 2012.
[20] A. Ghassemi, S. Bavarian, and L. Lampe, “Cognitive Radio for Smart Grid
Communications,” in IEEE International Conference on Smart Grid
Communications (SmartGrid Com), 2010, pp. 297-302.
[21] Z. Fadlullah, M. Fouda, N. Kato, A. Takeuchi, N. Iwasaki, and Y. Nozaki, “Toward
Intelligent Machine-to-Machine Communications in Smart Grid,” IEEE
Commun.Mag., vol. 49, no. 4, pp. 60-65, 2011.
[22] F. Huang, W. Wang, H. Luo, G. Yu, and Z. Zhang, “Prediction-based Spectrum
aggregation with hardware limitation in cognitive radio networks,” in Proceedings
of the IEEE 71st Vehicular Technology Conference (VTC '10), pp. 1–5, May 2010.
Page 2 - 1
CHAPTER 2
COGNITIVE MACHINE TO MACHINE
BACKGROUND
Page 2 - 2
2.1 Introduction
The use of M2M communications in the healthcare system has gained large momentum in the
past few years for some reasons, including distance monitoring of fitness information and
patient health, triggering alarms when vital conditions are discovered, and requesting
immediate treatment when needed. M2M sensors are usually deployed around the patient to
monitor fitness and health indicators like body temperature, blood pressure, weight, heart rate,
etc. These tracers typically follow certain protocols to communicate with each other and to
forward the data to gateways. Furthermore, these tracers (sensors) have energy thresholds and
are connected to the central monitoring system using a gateway or data aggregator by means
of short-range devices, the majority of which operate wirelessly [1].
This Chapter investigate and discuss cognitive machine-to-machine (CM2M) in e-healthcare
systems, the Chapter will explore a number of CR and M2M mechanisms, especially the ones
related to spectrum and energy efficiency in healthcare applications.
2.2 Machine to Machine in Healthcare Applications
M2M communications can enable automation in healthcare applications; which is important in
medical applications as it can improve the quality of healthcare by decreasing costs and enabling
continuous monitoring [2]. M2M automation will reduce the need for the active involvement
of medical personnel in gathering and analyzing patients’ data, and dispensing prescriptions.
Various types of sensors could be employed to collect and transfer the data, leading to faster
response time and preventing critical threats to a patient’s life. In [3] the classifications for
M2M in the healthcare sector has been proposed. A remote monitoring healthcare architecture
is presented [4, 5], where the authors covered the following classifications:
• c-Health - classical health care.
• e-Health - electronic health care, a subset of c-health.
• m-Health - mobile healthcare - the use of mobile devices, a subset of e-health.
• s-Health - smart healthcare - the use of a set of measures to deliver data and
enable the prevention of health hazards, e.g., informing on levels of pollution,
pollen, and allergens.
Page 2 - 3
The classification shows the needs for M2M communications in e-healthcare systems,
however, the study didn’t consider or address CR technology for better energy or spectrum
efficiency in M2M healthcare systems. In [6] the European Telecommunications Standards
Institute (ETSI) has marked a group of M2M healthcare applications such as aging
independently, disease handling, health betterment, and body fitness.
These marked applications can benefit from remote monitoring for a variety of conditions such
as cardiac arrhythmias and diabetes, which could support elderly people and make their lives
better. Furthermore, the usage of M2M in fitness machines and health facilities is discussed,
for example, the monitoring of breathing and heart rates, fat burning rate, and how much energy
is spent in a specific session or exercise.
Figure 2.1: ETSI System Architecture for M2M Networks in Healthcare Systems [6]
All this data could be forwarded immediately, safely and securely, to the servers, causing less
delay and faster communications between these machines and the end user. Remote healthcare
monitoring has been used more intensively over the past few years, identifying low body
Page 2 - 4
signals via secure communications between the patients and carers. ETSI in [6] defined a high-
level system architecture for M2M networks in healthcare systems. As shown in Figure 2.1,
the frequent necessity for hospitals to monitor patients for important indicators like blood and
oxygen pressure, electrocardiogram, and temperature has been discussed. Despite ETSI
provided us with a good system architecture for M2M networks in healthcare systems, their
architecture didn’t consider CR as smart technology to enable M2M communications in a
remote area with less spectrum and energy costs [7, 8].
Typically, M2M sensors connect with each other and with the gateways using wires but, due
to limitations and cost problems many vendors have adopted wireless technology as a
replacement for wire mediums in healthcare applications [9][10][11]. In [11], the use of
wireless technology instead of wires is proposed and investigated, as wireless technology is
promising to eliminate the use of wires. Enabling sensors to collect information from sensors
and send it back securely to end users will help to increase the efficiency and flexibility of
M2M medical applications.
Furthermore, using wireless communications can help with power efficiency and increase the
battery life of sensors. [10]. Also, wireless communications can enable the use of remote
monitoring services, reducing the cost of M2M sensors by using less expensive wireless
sensors to communicate with each other and forward data to the gateways. Global monitoring
can be extended to include a huge number of hospitals around the world by using such smart
and intelligent systems. Moreover, any change in a patient’s condition can be flagged up earlier,
enabling hospitals and healthcare facilities to respond more quickly.
However, M2M healthcare applications will face challenges with regard to Electromagnetic
Interferences (EMI) on the medical band, due to the massive number of devices. Also, mobile
operators will become busier and more expensive and would almost certainly become unable
to cope with the high volume of M2M traffic in the future [12]. To cope with these challenges
cognitive radio as intelligent technology could be exploited in healthcare applications for better
utilization of the medical band and an EMI-aware prioritized wireless access approach to avoid
the malfunction of healthcare devices [12].
Spectrum and energy efficiency are two of the main targets of M2M communications standards
[6]. Thus, it is very important to have a less crowded spectrum band and more energy efficient
communications in M2M medical applications. At present, bands like MedRadio (402–
Page 2 - 5
405 MHz) and Wireless Medical Telemetry Service (WMTS) are used in various types of
healthcare applications, but these types of bands are very limited in terms of spectrum
availability and have become very crowded recently [13].
Moreover, the Industrial, Scientific, and Medical (ISM) 2.4 GHz bands are not efficient for
important and critical medical applications due to the congestion and interference problem
caused by other IT wireless networks in hospitals and medical facilities such as normal Wi-fi
networks. By having other types of spectrum resource (e.g., cognitive radio spectrum resource),
the Quality of Service (QoS) for these important and critical applications can be improved [14].
In [14], the use of CR technology for sending medical and non-medical information to
pharmacies and medication centers is proposed. This type of information can communicate
wirelessly between the doctor and the nurse, and also offers video conferencing and
surveillance.
In [15] an energy efficient cognitive radio algorithm is proposed; the proposed algorithm
considers sending the information from the source (sensors) to the destination (servers) with
maximum energy efficiency within the CR sensors healthcare networks. The proposed
algorithm also considers the number of nodes and the distance between nodes. However, the
algorithm considered only the energy efficiency while ignoring the spectrum efficiency and
achieving user’s high data rate requirements.
In [16] The use of M2M applications to reduce obesity in high percentage obesity countries is
discussed. Mobile applications are used to recommend the types of food a person can have for
a specific day, based on the data received by the monitoring sensors. The authors in [10]
proposed M2M health monitoring and medication recommendation system for cardiovascular
patients who receive their treatments at home.
Patients’ critical signs (heart rhythm, blood pressure) will be supervised via M2M smart
sensors, which will send notifications to the M2M main server. Next, the main server will
assign medication recommendations automatically using techniques called artificial
intelligence, based on information garnered from medical records (hospitals) and other
healthcare facilities. Doctors will be involved in the process since the application deals with
critical diseases which may threaten a patient’s life. However, the system still complicated and
required high latency communications.
Page 2 - 6
In [5, 17] medical applications for M2M are studied and investigated. M2M will play a huge
part in medical applications, enabling the use of information technology without the need for
human control. But, generally, these applications need to be easy to install and cheap to buy
and will need to work with a good spectrum and energy efficiency.
2.3 Cognitive Radio Spectrum Sensing and Sharing
2.3.1 Spectrum Sensing
A key feature of CM2M communications is the ability to choose the best available channels by
exploiting cognitive ability and reconfigurability [18, 19]. As discussed at the beginning of this
Chapter, there is a limitation in the available spectrum, and the most significant challenge is to
share the primary user spectrum without affecting or interfering with the transmission of other
primary users (PU) to achieve primary user protection. CM2M devices typically utilize
temporarily unused spectrum, called white space or spectrum hole [19]. If this band is again
needed by the primary user, CM2M devices should either switch to other free channels or stay
in the current channel and stop transmission, changing its modulation scheme or transmission
power level to avoid interference with the licensed user.
CM2M devices can interact in real time with the available channels to choose the best
communication parameters and adapt to the dynamic radio environment. An illustration of the
CM2M devices activities is shown in Figure 2.2 [20, 21]. As shown in Figure 2.2 CM2M
devices using the free holes in the spectrum to make the transmission and improve the date of
the secondary users.
Page 2 - 7
Figure 2.2: CM2M Activities with Channel Transmitting [20]
2.3.2 Transmission Detection
In the CM2M transmission detection scheme, CM2M devices detect the weak signals of
licensed user transmissions through local observations [22, 23]. There are three approaches
under this scheme, which rely on the detection of the energy of a signal: matched filter [24,
25], cyclostationary feature detection [26, 27], and energy detection [28]. The first two
approaches are coherent detectors that achieve better detection probability but need information
about the licensed signal, while the third one (energy detector) is a non-coherent detector that
does not require knowledge about the licensed signal, and is cheaper and simpler to implement
compared to the other approaches.
1. Matched Filter Detection
This approach requires the licensed user to send a pilot signal with the data. The pilot signal
will be recognized by CM2M devices, enabling them to run timing and carrier synchronization
to attain coherence [25]. CM2M devices must have complete knowledge of the licensed user’s
signaling conditions, for example, operating frequency, bandwidth, and modulation type [24,
29]. The core benefit of this approach is a reduction in the time needed to provide a huge
processing gain, because of the coherent detector. Conversely, CM2M requires receivers for
Page 2 - 8
every type of primary system. Thus it increases the cost and the complexity and requires higher
energy consumption to detect several primary signals.
2. Cyclostationary Feature Detection
This is a method of detecting licensed user transmissions by utilizing the cyclostationary
characteristics of the received signals. The detection scheme can detect the random noise power
and so tell the difference between noise and the licensed user’s signals, thus achieving better
detection efficiency than the energy detector in discriminating against noise. However, this
approach is very expensive and complicated and needs a very long time for observation [30,
32].
3. Energy Detection
The energy detection approach is widely deployed for spectrum sensing due to its simplicity
and low cost. This approach is typically used when the licensed user signal is unknown [33].
The signal is detected by matching the output of the energy detector with desired constraints
which usually rely on the noise floor.
Typically, the effectiveness of the energy detection approach relies on the Signal-To-Noise
Ratio (SNR) value of the sender signal. In actual applications, the output signal at each CM2M
device may experience a problem from the hidden licensed user and suspicion due to
shadowing and fading problems. Cooperative sensing techniques addressed and solved some
of these issues [33, 34]. But, these techniques are based on the assumption that the noise power
levels are clearly known.
In practice, noise power levels change with the location and time of the station. This is called
noise uncertainty and cannot be predicted precisely. The effect of noise power levels
uncertainty on energy signal detection has recently been studied in [34, 35]. In [34], the primary
bounds of energy signal detection in the existence of noise uncertainty are investigated and
addressed. The investigation demonstrates that there are several SNR constraints under noise
uncertainty (e.g., 30 dB) that prevent satisfying, efficient detection. In [35], the authors suggest
a novel approach that utilises dynamic constraints to cope with the noise power fluctuation
issue, but as yet, the algorithm cannot ensure minimising spectrum sensing inaccuracy as more
sensing schemes need to be utilised to increase the sensing accuracy such as increasing the
number of antennas with the consideration of energy constraint.
Page 2 - 9
In [36-38], an optimal constraint based on the energy detection approaches have been
determined. The work also proves the efficiency, flexibility and the low cost that can be
achieved by using the energy detection scheme. However, the work didn’t consider the user
reliability and primary user protection for example high probability of detection, a collision
between the primary user and secondary user, and low probability of false alarm.
In conclusion, the behavior of energy detection depends on the accuracy of detection
constraints and the SNR level of the output signal. From a CM2M perspective, the energy
detection scheme is the best scheme that can be utilized to work with CM2M communications
for its simplicity and low cost of the implementation. But yet more work needs to be developed
to address the problems of user reliability and primary users protection.
2.3.3 Cooperative detection
The cooperative detection means, the cooperation among CM2M devices to reduce the
uncertainty caused by the single user’s detection. Using a number of sensing nodes, cooperative
sensing can be utilized to mitigate multipath fading and spatial diversity, which are the key
points that decrease the performance of single user’s detection. Cooperative detection can
achieve more precise performance [39].
However, it is more complicated and needs more overhead traffic and operations to cooperate
with other CM2M devices. As a result, employing cooperative detection in CM2M networks
will add more cost and complexity to its parameters, CM2M generally could utilize a cheaper
and less complex detection such as energy detection to sense and find an available channel in
the CM2M radio environment [35, 40].
2.3.4 Spectrum sharing
Based on the access schemes, spectrum sharing can be divided into [41]
1. Overlay Spectrum Sharing: In this kind of spectrum sharing scheme, CM2M users
can access part of the spectrum that is not being used by the primary user. Consequently,
there will be less interference between primary users and secondary users.
2. Underlay Spectrum Sharing: This kind of spectrum access scheme utilizes the spread
spectrum mechanisms created for mobile operators. CM2M users start communication
at the particular part of the spectrum assigned by a spectrum map with transmission
Page 2 - 10
energy regarded as noise by the primary user. This kind of scheme can be used to
maximize the bandwidth compared to the overlay scheme.
2.4 Spectrum Aggregation
The massive wireless volume generated by the modern era of machines (wearable sensors,
smartphones, M2M devices, laptops..etc.) will soon become too large for the current networks
to handle [42, 20]. The exploitation of extra spectrum resources with better throughput rate is
one of the significant solutions for high demand and traffic explosion. Generally, because of
the restricted spectrum allocation policy, wireless networks can only employ a continuous
spectrum resource. Moreover, the wide continuous spectrum bands are difficult to utilize under
the existing situation of spectrum resource. In the past few years, a huge amount of work both
in industry and academia has encouraged the use of the flexible spectrum option to combat the
so-called spectrum fragmentation problem.
To cope with such a problem, Spectrum Aggregation (SA) is proposed to boost system
efficiency by adding extra bandwidth for mobile users. SA utilizes fragmented spectrum
segments to expand bandwidth transmission. After CR finds white space bands sensed by its
smart inspecting capabilities, the combination of diverse spectrum bands becomes optional by
utilizing SA. A number of research centers, companies, and standard organizations began to
address and study SA such is QinetiQ, which has developed a novel solution for SA to sustain
broadband services [43]. CR is promising technology but comes with some challenges, one of
which is the spectrum assignment problem [44, 45].
The spectrum assignment problem is extensively addressed and studied in typical wireless
networks [35]. When considering spectrum assignation, the goal is to achieve the interference
thresholds and increase the system data rate for the given spectrum. Moreover, CM2M devices
could cause a number of challenges due to the discontinuous and fluctuating nature of the
available spectrum, in addition to the various QoS needs of different applications. Despite the
number of standardizations addressing SA in Long-Term Evolution-Advance (LTE-A)
networks, little work has been done to utilize SA by exploiting CR technology. There is very
limited literature considering spectrum assignment among users having SA capabilities.
Page 2 - 11
In [46], proposed a prediction based spectrum aggregation scheme to increase capacity and
decrease the re-allocation overhead. The proposed scheme is referred to as a Maximum
Satisfaction Algorithm (MSA) for spectrum assignment. The main idea is to assign spectrum
for the user with the largest bandwidth requirement first, leaving better spectrum bands for
remaining users, while taking into consideration the different bandwidth requirements of users
and channel state statistics.
Later, Fang in [47] introduced a genetic algorithm based spectrum assignment in CR networks.
Authors in [48], suggest a utility-based SA algorithm to boost the performance of CM2M
devices considering a number of objectives: 1) decrease the amount of channel switching. 2)
increase the overall data rate. 3) decrease the number of sub-channels comprising the aggregate
channel, aimed at opportunistic spectrum use by CM2M devices. The objectives work
simultaneously with a weighted sum utility function. Furthermore, the suggested algorithm
allows for the automated adaptable setting of objective-function weights depending on
available channels.
In [49, 50], an Aggregation Aware Spectrum Assignment Algorithm (AASAA) is proposed to
aggregate discrete spectrum fragments hugely. The algorithm in [50] utilizes the first available
aggregation range from the low-frequency requirement. In [51], the authors analytically
addressed the channel assignment optimization problem and the channel access problem as
joint power control, with the goal of decreasing the required spectrum needs for specific CR
parameters.
The optimization problem in [52] appeared to be a binary linear program, which is, usually a
Non-deterministic Polynomial-time (NP)-hard problem. Thus, [53] suggest a near-optimal
solution according to sequential fixing, where the binary variables are calculated iteratively by
addressing a number of linear programs. For CM2M devices, the current SA and spectrum
assignment solutions are not applicable directly as practical thresholds, such as MAS, must be
taken into consideration. Furthermore, in an aggregation-based spectrum assignment, a
significant challenge is to manage CCI among CM2M devices, again not taken into
consideration in the current literature.
Page 2 - 12
2.5 Spectrum Sensing and Sharing
A number of studies addressed and considered energy and spectrum efficiency sensing and
sharing schemes for better communications in CM2M networks. In [9] energy efficient
spectrum discovery design in the smart grid is considered. The design aims to improve the
energy efficiency of the CM2M networks using machine coordination and assumes that
machines cooperating with each other will decrease energy costs during the spectrum discovery
phase. The results were promising, but the design has not considered the main requirements of
CM2M networking relating to cost and complexity. The design considers cooperative sensing
for improved sensing accuracy but has not taken into account the additional cost and
complexity that cooperative sensing adds to the design.
In [45], CR in M2M communications has been addressed from a protocol stack perspective.
However, the work needs more reliable and effective implementation of cognitive M2M
networks. In [55], Weightless is proposed as an open standard for exchanging data between a
base station and a huge number of devices around it. Weightless allows developers to build low
energy wide area networks with a high throughput data rate by exploiting CR technology.
Weightless aims and objectives are: to prolong battery life, reduce installation costs, allow wide
outdoor and indoor coverage, and avoid interference caused by another unlicensed user.
Weightless is still under review by 3GPP and other communications parties, achieving its
targets will need more time and new schemes and algorithms especially, with the ones related
to the energy efficiency and coverage.
The authors in [56] studied decreasing energy cost per bit for distributed cognitive radio
devices in different types of network; the study assumed joint source and channel sensing for
CR devices should be analyzed first, and then the optimal solution could be determined from
fading channels. In [57] energy efficiency in CR wireless devices is addressed, the work of
distributed sensing approach optimizes power efficiency with thresholds on the minimum
desired detection probability, and the increase of the permissible probability of false alarms by
selecting efficient sensing and sleeping design energy parameters proposed to improve energy
efficiency. However, both studies didn’t consider spectrum sharing and sleep mode techniques
that could be exploited to achieve more energy efficiency by switching the SUs to off mode
when no data need to be transmitted.
Page 2 - 13
In [58] the authors present two kinds of technique, real-time and non-real-time, which can split
the SU into two priorities and develop a spectrum handoff scheme and a resource reservation
mechanism to assure the QoS of SUs. In [59], a different type of opportunistic spectrum access
approach is explored to provide the primary user and the secondary user with better QoS. Both
users are analyzed by pre-emptive and non-pre-emptive priorities with the basic queueing
system. In [60] the authors present a channel allocation system to enhance the QoS of the
priority based secondary user with improved performance in terms of data rate, dropping
probability, and blocking. Furthermore, the developed analytical Markov systems considerably
enhance the multimedia QoS performance for the critical priority SUs.
However, the above studies neglected the energy efficiency of SU while the focus was only on
the QoS for SUs, in e-healthcare CM2M applications it’s very important to have energy
efficient communications to prolongs SU battery life and meet M2M expectations.
2.6 Learning and Channel Selecting
A huge amount of work has been done regarding learning algorithms for cognitive radio. Such
as genetic systems, neural models, and the algorithms of Markova [60]. Galindo-Serrano, L.
Giupponi, [61] suggest a frame composed of real-time decentralized Q-learning to control
Wireless Regional Area Network (WRAN) systems and aggregated interference.
Katidiotis, [62] exploit neural models for learning to predict data bit rate for CR networks (e.g.,
CM2M gateways). Yang, M. F. and D. Grace [63] suggest intelligence based on reinforcement
learning and transmitter energy system for better channel assignment in terrestrial multicast
systems. Torkestani, [64] developed the learning based CR to solve the problem of the
spectrum limitation in wireless ad hoc networks.
Panahi and Ohtsuki [65] present a Fuzzy Q Learning (FQL) based scheme for channel sensing
in CR networks. Zhang in [66] presented a reinforcement learning-based double action
algorithm intended to increase the efficiency of dynamic spectrum access in CR networks. A
number of studies employ partially observable Markov decision processes [66,67] and
reinforcement learning [65, 68]. The main disadvantage of these techniques is their dependence
on highly accurate reward functions.
In [67] Meybodi and Torkestani suggest learning automation in CR to cope with the problems
of spectrum limitations in wireless ad hoc networks. In [69] Zhu developed a reinforcement
Page 2 - 14
learning scheme to find an efficient protocol under undiscovered environment. The above
proposals are comparatively useful when it comes to practical applications for huge wireless
systems. But by taking into consideration, a new number of studies have concentrated on
Experience Weighted Attraction (EWA) intelligent channel handoff approaches, due to
stability and sensitivity advantages, conducted by comparing Q learning. A little effort has
taken place with regard to designing and developing a learning engine for CR with (EWA)
algorithms.
EWA algorithms [70, 71] give CR the ability to be aware of available channels characteristics
online. By collecting the history of channel statuses, it can foresee, select, and amend the best
available channel, dynamically test the quality of communication links, and eventually
decrease system communication outage probability. The efficiency of this algorithm has been
tested by the straightforward probability approach [72] and with the EWA Handoff (EWAH)
algorithm [69] in our preliminary studies.
2.7 Summary
The present situation and the development of M2M and CR communications in healthcare
applications are investigated and discussed in this Chapter. Furthermore, the viability of using
cognitive radio as smart technology to achieve better energy and spectrum efficiency is
explored. Current efforts utilize cognitive radio technology to prolong the battery life of sensors
and improve sensing accuracy. In addition, the current efforts exploit CR radio for better
spectrum utilization and throughput efficiency. However, for CM2M devices, the current SA
and spectrum assignment solutions are not applicable directly as practical thresholds, such as
MAS, must be taken into consideration.
Furthermore, in aggregation based spectrum assignment a significant challenge is to manage
Co-Channel Interference (CCI) among CM2M devices, which is not taken into consideration
in the current literature. On the other hand, no previous work has considered both spectrum
handoffs and the wait/switch trade-off in a CM2M network with multiple CM2M devices (e.g.,
CM2M gateways) for more spectrum and energy efficient communications. Moreover, no
previous work considers designing and developing a learning engine for CR with (EWA)
Page 2 - 15
algorithms for better spectrum and energy efficiency in CM2M communications. In the next
Chapters, this thesis will consider the methods above/techniques as solutions to achieve better
spectrum and energy efficiency in healthcare CM2M communications.
Page 2 - 16
References for Chapter 2
[1] Zhong Fan and S. Tan, "M2M communications for e-health: Standards, enabling
technologies, and research challenges," 2012 6th International Symposium on
Medical Information and Communication Technology (ISMICT), La Jolla, CA,
2012, pp. 1-4.
[2] Yang Y, Ye H & Fei S (2011) Design of communication interface for M2M-based
positioning and monitoring system. Proceedings of International Conference on
Electronics, Communications, and Control, Ningbo, China: 2624–2627.
[3] Solanas A., at all: A Context-Aware Health Paradigm within Smart Cities, IEEE
Comm. Magazine, 08.2014, pp. 74-814.
[4] ISO/IEEE 11073, Health informatics - Point-of-care medical device
communication, 10101-Nomenclature, 10201-Domain information model, 30200-
Transport profile - Cable connected, 30300-Transport profile - Infrared Wireless.
11073-10404 to 20601.
[5] J. Jarmakiewicz, K. Parobczak, and K. Maślanka, "On the Internet of Nano Things
in healthcare network," 2016 International Conference on Military Communications
and Information Systems (ICMCIS), Brussels, 2016, pp. 1-6.
[6] ETSI, Machine to Machine Communications (M2M): Use cases of M2M
applications for eHealth, ETSI TR 102 732, 2011.
[7] I.F. Akyildiz, Won-Yeol Lee, Mehmet C. Vuran, and S. Mohanty. A survey on
spectrum management in cognitive radio networks. Communications Magazine,
IEEE, 46(4):40{48, April 2008. ISSN 0163-6804.
[8] Illanko, K., Naeem, M., Anpalagan, A., & Androutsos, D. (2011). Energy-Efficient
Frequency and Power Allocation for Cognitive Radios in Television Systems. IEEE
Systems Journal, 10(1), 313-324.
[9] Quoc Duy Vo, Joo-Pyoung Choi “Green Perspective Cognitive Radio-based M2M
Communications for Smart Meters”, Tutorials, IEEE 978-1-4244-98072010.
[10] W. Zhao, W. Chaowei, and Y. Nakahira, “Medical application on the Internet of
Things,” in Proc. IET Int. Conf. Commun. Technol. Appl. (ICCTA), Oct. 2011, pp.
660-665.
[11] J. Wang, M. Ghosh and K. Challapali, "Emerging cognitive radio applications: A
survey," in IEEE Communications Magazine, vol. 49, no. 3, pp. 74-81, March 2011.
Page 2 - 17
[12] I. A. Mamoon, A. K. M. Muzahidul-Islam, S. Baharun, S. Komaki and A. Ahmed,
"Architecture and communication protocols for cognitive radio network enabled
hospital," 2015 9th International Symposium on Medical Information and
Communication Technology (ISMICT), Kamakura, 2015, pp. 170-174.
[13] M. Patel and J. Wang, "Applications, challenges, and perspective in emerging body
area networking technologies," in IEEE Wireless Communications, vol. 17, no. 1,
pp. 80-88, February 2010.
[14] I. A. Mamoon, A. K. M. Muzahidul-Islam, S. Baharun, S. Komaki and A. Ahmed,
"Architecture and communication protocols for cognitive radio network enabled
hospital," 2015 9th International Symposium on Medical Information and
Communication Technology (ISMICT), Kamakura, 2015, pp. 170-174.
[15] A. Semwal, H. S. Bhadauria, and A. Singh, "Energy efficient approach to sending
data in cognitive radio wireless sensor networks(CRSN)," 2016 International
Conference on Advances in Computing, Communications and Informatics
(ICACCI), Jaipur, 2016, pp. 2047-2051.
[16] G. Wibisono and I. G. B. Astawa, "Designing Machine-to-Machine (M2M)
Prototype System for Weight Loss Program for Obesity and Overweight Patients,"
2016 7th International Conference on Intelligent Systems, Modelling and
Simulation (ISMS), Bangkok, 2016, pp. 138-143.
[17] Chung WY, Yau C, Shin KS & Myllyla, R (2007) A cell phone based health
monitoring system with self-analysis processor using wireless sensor network
technology. Proceedings of the 29th Annual International Conference of the IEEE
EMBS, Lyon, France: 3705–3708.
[18] F. J. Lin, Y. Ren and E. Cerritos, "A Feasibility Study on Developing IoT/M2M
Applications over ETSI M2M Architecture," 2013 International Conference on
Parallel and Distributed Systems, Seoul, 2013, pp. 558-563.
[19] Van-Tam Nguyen,1 Frederic Villain,2 and Yann Le Guillou, “Cognitive Radio RF:
Overview and Challenges,” VLSI Design, Volume 2012 (2012), Article ID 716476,
13 pages.
[20] Junfeng Xiao, R.Q. Hu, Yi Qian, Lei Gong, and Bo Wang. Expanding LTE network
spectrum with cognitive radios: From concept to implementation. Wireless
Communications, IEEE, 20(2):12{19, April 2013. ISSN 1536-1284.
[21] Baronti P, Pillai P, Chook VWC, Chessa S, Gotta A & Hu YF (2007) Wireless
sensor networks: A survey on state of the art and the 802.15.4 and ZigBee standards.
Computer Communications 30: 1655–1695.
Page 2 - 18
[22] I. F. Akyildiz, W. y. Lee, M. C. Vuran and S. Mohanty, "A survey on spectrum
management in cognitive radio networks," in IEEE Communications Magazine, vol.
46, no. 4, pp. 40-48, April 2008.
[23] Li, G.; Xu, Z.; Xiong, C.; Yang, C.; Zhang, S.; Chen, Y.; Xu, S. Energy-efficient
wireless communications: tutorial, survey, and open issues IEEE Wirel. Commun.
2011, 18, 28–35.
[24] Z. Quan, S. Cui, H. Poor, and A. Sayed, “Collaborative wideband sensing for
cognitive radios,” Signal Processing Magazine, IEEE, vol. 25, no. 6, pp. 6073,2008.
[25] N. L. Johnson, S. Kotz, and N. Balakrishnan, Continuous Multivariate Distributions,
volume 1, Models, and Applications. New York: John Wiley & Sons, 2002.
[26] P. Dharmawansa, N. Rajatheva, and K. AHMED, “On the distribution of the sum of
nakagami-m random variables,” IEEE transactions on communications, vol. 55, no.
7, pp. 1407–1416, 2007.
[27] S. P. Herath and N. Rajatheva, “Analysis of equal gain combining in energy
detection for cognitive radio over nakagami channels,” in Global
Telecommunications Conference, 2008. IEEE GLOBECOM 2008. IEEE. IEEE,
2008, pp. 1–5.
[28] R. R. Tenney and N. R. Sandell, “Detection with distributed sensors,” Aerospace
and Electronic Systems, IEEE Transactions on, no. 4, pp. 501–510, 198.
[29] M. Mustonen, M. Matinmikko, and A. Mammela, “Cooperative spectrum sensing
using quantized soft decision combining,” in Cognitive Radio Oriented Wireless
Networks and Communications, 2009. CROWNCOM’09. 4th International Con-
ference on. IEEE, 2009, pp. 1–5.
[30] G. Huang and J. K. Tugnait, "On Cyclostationarity Based Spectrum Sensing Under
Uncertain Gaussian Noise," in IEEE Transactions on Signal Processing, vol. 61, no.
8, pp. 2042-2054, April 15, 2013. doi: 10.1109/TSP.2013.2246158.
[31] K. Ben Letaief and W. Zhang, “Cooperative communications for cognitive radio
networks,” Proceedings of the IEEE, vol. 97, no. 5, pp. 878–893, 2009.
[32] T. Yucek and H. Arslan, "A survey of spectrum sensing algorithms for cognitive
radio applications," IEEE Communications Surveys & Tutorials, vol. 11, pp. 116-
130, 2009.
[33] K. Ben Letaief and Z. Wei, "Cooperative Communications for Cognitive Radio
Networks," Proceedings of the IEEE, vol. 97, pp. 878-893, 2009.
Page 2 - 19
[34] R. Tandra and A. Sahai, "Fundamental limits on detection in low SNR under noise
uncertainty," in International Conference on Wireless Networks, Communications
and Mobile Computing, 2005, pp. 464-469.
[35] R. Tandra and A. Sahai, "SNR Walls for Signal Detection," IEEE Journal of
Selected Topics in Signal Processing, vol. 2, pp. 4-17, 2008.
[36] Wnbo, et al., "The Performance Merit of Dynamic Threshold Energy Detection
Algorithm in Cognitive Radio Systems," 1st International Conference on
Information Science and Engineering (ICISE), pp. 692-695, 2009.
[37] X. Shujing, et al., "Optimal threshold of energy detection for spectrum sensing in
cognitive radio," International Conference on Wireless Communications & Signal
Processing, WCSP '09. , pp. 1-5, 2009.
[38] Z. Wei, et al., "Cooperative Spectrum Sensing Optimization in Cognitive Radio
Networks," IEEE International Conference on Communications, ICC '08., pp. 3411-
3415, 2008.
[39] C. Liu, M. Li, and M.-L. Jin, “Blind energy-based detection for spatial spectrum
sensing,” IEEE Wireless Commun. Lett., vol. 4, no. 1, pp. 98–101, Feb. 2015.
[40] S. Atapattu, C. Tellambura, and H. Jiang, "Energy Detection Based Cooperative
Spectrum Sensing in Cognitive Radio Networks," Wireless Communications, IEEE
Transactions on, vol. 10, pp. 1232-1241, April 2011.
[41] I. Akyildiz, W. Lee, M. Vuran, and S. Mohanty, “Next generation/dynamic
spectrum access/cognitive radio wireless networks: a survey,” Computer Networks,
vol. 50, no. 13, pp. 2127–2159, 2006
[42] Cisco visual networking index: Global mobile data forecast update 2014{2019 white
paper, 2015. URL http://www.cisco.com/c/en/us/solutions/collateral/service-
provider/ visual-networking-index-vni/white_paper_c11-520862.html.
[43] Ltd. QinetiQ. A study of the provision of aggregation of frequency to provide wider
bandwidth services. http://www.aegis-systems.co.uk/download/ 1722/aggregation
.pdf/, Aug. 2006. [report for Oce of Communications(Ofcom), UK, Online; accessed
07-May-2013].
[44] I.F. Akyildiz, Won-Yeol Lee, Mehmet C. Vuran, and S. Mohanty. A survey on
spectrum management in cognitive radio networks. Communications Magazine,
IEEE, 46(4):4048, April 2008. ISSN 0163-6804.
Page 2 - 20
[45] User equipment (ue) radio access capabilities (release 12). Technical report, 3GPP
TS 36.306, MAY. 2014. URL http://www.3gpp.org.
[46] Furong Huang, Wei Wang, Haiyan Luo, Guanding Yu, and Zhaoyang Zhang.
Prediction-based spectrum aggregation with hardware limitation in cognitive radio
networks. In Vehicular Technology Conference (VTC 2010-Spring), 2010 IEEE
71st, pages 15, 2010.
[47] Fang Ye, Rui Yang, and Yibing Li. Genetic algorithm based spectrum assignment
model in cognitive radio networks. In Information Engineering and Computer
Science (ICIECS), 2010 2nd International Conference on, pages 14, Dec 2010.
[48] Haeyoung Lee, S. Vahid, and K. Moessner. Utility-based dynamic spectrum
aggregation algorithm in cognitive radio networks. In Vehicular Technology
Conference (VTC Fall), 2012 IEEE, pages 15, Sept 2012. doi: 10.1109/
VTCFall.2012.6399312.
[49] Dawei Chen, Qian Zhang, and Weijia Jia. Aggregation aware spectrum assignment
in ad-hoc cognitive networks. In Cognitive Radio Oriented Wireless Networks and
Communications, 2008. CrownCom 2008. 3rd International Conference on, pages
16, 2008.
[50] D. Chen, Q. Zhang, and. Jia, “Aggregation aware spectrum assignment in ad-hoc
cognitive networks,” in Cognitive Radio Oriented Wireless Networks and
Communications, 2008. CrownCom 2008. 3rd International Conference on, 2008,
pp. 1–6.
[51] H.A.B. Salameh, M. Krunz, and D. Manzi. Spectrum bonding and aggregation with
guard-band awareness in cognitive radio networks. Mobile Com- puting, IEEE
Transactions on, 13(3):569{581, March 2014. ISSN 1536-1233.
[52] C. Wijting et al. Key Technologies for IMT-Advanced Mobile Communication
Systems. Wireless Communications, IEEE, 16(3):76{85, June 2009. ISSN 1536-
1284. doi: 10.1109/MWC.2009.5109467.
[53] Li-Chun Wang and Chung-Wei Wang. Spectrum management techniques with QoS
provisioning in cognitive radio networks. In Wireless Pervasive Computing
(ISWPC), 2010 5th IEEE International Symposium on, pages 116{121, May 2010.
doi: 10.1109/ISWPC.2010.5483800.
[54] Aijaz, A., & Aghvami, A.H. (2015). Cognitive Machine-to-Machine
Communications for Internet-of-Things: A Protocol Stack Perspective. Internet of
Things Journal, 2(2), 103-112.
Page 2 - 21
[55] Weightless, http://www.weightless.org/about/what-is-weightless. (Accessed: 29th-
June 2016).
[56] Gao S., Qian L., Vaman D. Distributed Energy-Efficient Spectrum Access in
Wireless Cognitive Radio Sensor Networks. Proceedings of 2008 Wireless
Communications and Networking Conference; Las Vegas, NV, USA. 31 March–3
April 2008; pp. 1442–1447.
[57] Hu W., Dinh T.L., Corke P., Jha S. Outdoor sensor net design, and deployment:
Experiences from a Sugar Farm. IEEE Pervasive Computer. 2012;11:82–91.
[58] Jinsheng Yang and Xinming Shao, "Optimal number of secondary users in weighted
cooperative spectrum sensing," Proceedings of 2011 Cross Strait Quad-Regional
Radio Science and Wireless Technology Conference, Harbin, 2011, pp. 902-905.
[59] F. Zhou, and G. Liu, “Quality of hybrid services in cognitive radio networks,”
Consumer Electronics, Communications and Networks (CECNet), 2013 3rd
International Conference on IEEE., pp. 457–460, Nov. 2013.
[60] N. Khedun, and V. Bassoo, “Analysis of priority queueing with multichannel in
cognitive radio network” EUROCON 2015-International Conference on Computer
as a Tool (EUROCON), IEEE., pp. 1–6, Sep. 2015.
[61] Q-Learning for Aggregated Interference Control in Cognitive Radio Networks,
IEEE Transactions on Vehicular Technology, Vol. 59, Issue 4, 2010, pp. 1823-1834
[62] A. He, J. Gaeddert, K. K. Bae, T. R. Newman, J. H. Reed, L. Morales, and C.-H.
Park, “Development of a case-based reasoning cognitive engine for IEEE 802.22
warn applications,”.
[63] Yang, M. F. and D. Grace, Cognitive Radio with Reinforcement Learning Applied
to Multicast Downlink Transmission with Power Adjustment, Wireless Personal
Communications, Vol. 57, Issue 1, 2011, pp. 73-87.
[64] Torkestani, J. A. and M. R. Meybodi, A Learning Automata-Based Cognitive Radio
for Clustered Wireless Ad-Hoc Networks, Journal of Network and Systems
Management, Vol. 19, Issue 2, 2011, pp. 278-297.
[65] Li, H. S., Multiagent Q-Learning for Aloha-Like Spectrum Access in Cognitive
Radio Systems, Eurasip Journal on Wireless Communications and Networking,
2010, pp. 1-15.
Page 2 - 22
[66] Shan-Shan, W., et al., Primary User Emulation Attacks Analysis for Cognitive
Radio Networks Communication, TELKOMNIKA Indonesian Journal of Electrical
Engineering, Vol. 11, Issue 7, 2013, pp. 3905-3914.
[67] Chen, X. F., et al., Reinforcement Learning Enhanced Iterative Power Allocation in
Stochastic Cognitive Wireless Mesh Networks, Wireless Personal Communications,
Vol. 57, Issue 1, 2011, pp. 89-104.
[68] Khalaf, G., An Optimal Sensing Algorithm for Multiband Cognitive Radio
Network, International Journal of Information and Network Security (IJINS), Vol.
2, Issue 1, 2013, pp. 60-67.
[69] Zhang, W. Z. and X. C. Liu, Centralized Dynamic Spectrum Allocation in Cognitive
Radio Networks Based on Fuzzy Logic and Q-Learning. China Communications,
Vol. 8, Issue 7, 2011, pp. 46-54.
[70] Zhu, J., et al., Adaptive transmission scheduling over fading channels for energy-
efficient cognitive radio networks by reinforcement learning, Telecommunication
Systems, Vol. 42, Issue 1-2, 2009, pp. 123-138.
[71] Y. Sun and J.-S. Qian, “Cognitive radio channel selection strategy based on
experience-weighted attraction learning,” TELKOMNIKA Indonesian Journal of
Electrical Engineering, vol. 12, no. 1, pp. 149–156, 2014.
[72] Gallego, J. R., M. Canales, and J. Ortin, Distributed resource allocation in cognitive
radio networks with a game learning approach to improve aggregate system
capacity, Ad Hoc Networks, Vol. 10, Issue 6, 2012, pp. 1076-1089.
Page 3 - 1
CHAPTER 3
AGGREGATION BASED SPECTRUM
ASSIGNMENT FOR COGNITIVE M2M
NETWORK
Page 3 - 2
3.1 Introduction
According to the Cisco company, a single M2M device currently produces as much traffic as
three basic-feature phones while due to emerging applications and services of M2M networks
the average traffic per device is forecasted to grow from 70 MB per month in 2014 to 366 MB
per month in 2018 due to QoS requirements [1]. Because of the high growth rate of the number
of devices and high demand for data traffic, the next generations of M2M networks will face a
number of challenges, especially, with the spectrum limitations problem. Cognitive Radio (CR)
comes as a promising technology to cope with the spectrum limitations problem in M2M
networks.
CR has become one of the most intensively studied paradigms in wireless communications,
allowing unlicensed users (e.g., CM2M devices or gateways) to opportunistically access
licensed spectrum as long as interference to PUs is kept at an acceptable level [2]. A number
of M2M applications (e.g., healthcare applications) can benefit from the combination of CR
and M2M communications [3]. CM2M networks can increase spectrum utilization and energy
efficiency in M2M networks [4].
The CM2M device can interact with available channels either by performing spectrum sensing
and accessing spectrum databases to detect available channels [4]. After sensing, the CM2M
device utilises the discovered unused spectrum according to the device needs, for example, TV
bands which have significantly favourable propagation features are typically booked to
broadcasters but after the recent transition from the analogue broadcast television system to the
digital one, a massive number of TV channels (also known as TV White Spaces) were freed
up and became unused.
In September 2010, the Federal Communications Commission (FCC) released an important
ruling [5], enabling unlicensed broadband wireless devices to utilize TV White Space (TVWS).
Unfortunately, due to spectrum fragmentation and as a result of an inefficient command and
control spectrum management scheme, a continuous wide segment of TVWS is rare in many
countries including the United Kingdom. As CM2M networks can sense and be aware of their
radio environment, the aggregation of narrow spectrum opportunities becomes possible.
Spectrum aggregation provides wider bandwidth and higher throughput for the CM2M
devices. CM2M devices can access discontinuous portions of the TVWS simultaneously using
Page 3 - 3
Discontinuous Orthogonal Frequency Division Multiplexing (DOFDM) [6, 7]. DOFDM is a
multi-carrier modulation technique and is a variant of OFDM used to aggregate discontinuous
segments of spectrum. The main difference between OFDM and DOFDM is the ON/OFF
subcarrier information block [8]. Multiple segments of the spectrum can be occupied by CM2M
devices or PUs. As a result, these subcarriers are off-limits to CM2M devices [7]. Thus, to
avoid interfering with other transmissions, the subcarriers within their vicinity are turned off
and become unusable for CM2M devices, as shown in Figure 3.1.
Furthermore, active (usable) subcarriers are located in the unused segments of spectrum, which
are calculated by spectrum aggregation to be one of the most important LTE Advanced
technologies from the physical layer perspective and standardized in LTE Release 10 [9].
Figure 3.1: Subcarrier Distribution Over Spectrum [79]
However, despite the standardization of spectrum aggregation, little effort has been made to
optimize spectrum aggregation by exploiting CR technology in M2M networks, as discussed
in Chapter 2. There is limited literature available on spectrum assignment among CM2M
devices having spectrum aggregation capabilities. In addition, for CM2M networks, existing
spectrum assignment and aggregation solutions are not applicable directly to practical issues
such as Maximum Aggregation Span (MAS) must be taken into account. In aggregation-based
spectrum assignment, a major challenge is to manage Co-Channel Interference among CM2M
devices which are not taken into account in the existing literature. The major contributions of
this chapter are listed below:
1. To prevent multiple CM2M devices from colliding in the overlapping portions of the
spectrum, a centralized approach is applied. Furthermore, an integer optimization
Page 3 - 4
mechanism is formulated to maximize cell-throughput, considering CCI and MAS in
an aggregation-aware CM2M network.
2. As the spectrum assignment problem is inherently seen as a Non-deterministic
Polynomial (NP) - hard optimization problem, evolutionary approaches can be applied
to solve this challenging problem. In this Chapter, GA is used to solve the aggregation
aware spectrum assignment because of its simplicity, robustness and fast convergence
of the algorithm [10].
The rest of this Chapter is organized as follows: In Section (3.2), the spectrum assignment and
aggregation models are presented. The proposed algorithm is explained in Section (3.3).
Simulation results are discussed in Section (3.4), followed by the conclusion in (3.5).
3.2 Spectrum efficiency (Systems Model)
3.2.1 Spectrum Assignment Model
A spectrum assignment model presumes to work with a CM2M network consisting of N
CM2M devices (𝝓) defined as:
𝝓 = {ɸ1, ɸ2, … . . ɸN}
competing for M non-overlapping orthogonal channels ( Y ) in uplink
Y = { y1, y2 … … . y𝑀 }
All spectrum assignment and access procedures are managed by a central entity called a
spectrum broker. Furthermore, the model assumes a distributed sensing scheme and
measurement conducted by each device is forwarded to the spectrum broker [11]. A spectrum
occupancy map is constructed at the spectrum broker, and CCI among CM2M devices is
Page 3 - 5
calculated. In addition, the spectrum broker can lease single or multiple channels for ɸn ∈ 𝝓
in a fixed geographical region for a duration of time. Ultimately, a base station can transmit
data to CM2M devices ɸn in the selected channels. Figure 3.2 depicts the system model used
in this Chapter, as shown in the Figure the spectrum broker assign (Y) channels to the CM2M
devices in order to make the transmission among the other remote CM2M devices using the
available specrum in the radio environment.
Figure 3.2: Architecture Diagram of CM2M network Operating in
TVWS
The channel availability defined by matrix (𝑳𝑳) anad as shown below
𝑳𝑳 = { 𝐿𝑛,𝑚,| 𝐿𝑛,𝑚, ∈ {0,1}}𝑁𝑥𝑀
Page 3 - 6
as a 𝑁 𝑥 𝑀 binary matrix representing channel availability where 𝐿𝑛,𝑚 = 1 if and only if
channel y𝑚 is available to CM2M ɸn and 𝐿𝑁,𝑀, = 0 otherwise, each CM2M ɸn associated
with a set of available channels at its location is defined as:
��𝑛 ⊂ 𝑌 ; i. e ��𝑛 = { y𝑚 | 𝐿𝑛,𝑚 ≠ 0}
Due to the different interference range of each PU (which depends on the PU’s transmission
power and the physical distance) at the location of each CM2M device, Y𝑛 of different CM2M
devices may be different [12, 13]. According to the sharing agreement, any y𝑚 ∈ Y can be
reused by a group of CM2M devices in the vicinity defined by 𝝓 𝒎 such 𝝓 𝒎 ⊂ 𝝓 if CM2M
devices are located outside the interference range of PUs; i.e. 𝝓 𝒎 = {ɸn|𝐿𝑛,𝑚 ≠ 0}.
The interference constraint matrix defined as (𝑨𝑨) and as show below
𝑨𝑨 = { 𝐴𝑛,𝑘,𝑚 | 𝐴𝑛,𝑘,𝑚 ∈ {0,1} }𝑁𝑥𝑁𝑥𝑀
N x N x M binary matrix representing the interference threshold among CM2M devices where
𝐴𝑛,𝑘,𝑚 = 1 if CM2M ɸn and ɸk would interfere with each other on y𝑚 , and 𝐴𝑛,𝑘,𝑚 = 0
otherwise. It must be noted from 𝑛 = 𝑘, 𝐴𝑛,𝑘,𝑚 = 1 − 𝐿𝑛,𝑚 . The value of 𝐴𝑛,𝑘,𝑚 = 1
depends on the distance between ɸn and ɸk.
Interference thresholds also depend on y𝑚 as power and transmission rules change massively
in various frequency bands. The bandwidth requirements of all CM2M devices are different
because of the various quality of service requirements for each device. The model defines 𝑅 =
{𝑟𝑛}1𝑥𝑁 as device request bandwidth, where 𝑟𝑛 represents a bandwidth demand of ɸn.
In a dynamic environment, the availability of channels and the interference threshold matrix
both change continually; the spectrum availability is presumed to be static or varies slowly in
each scheduling time slot (e.g., all matrices remain constant through the scheduling duration
Page 3 - 7
time). Furthermore, the model presumes a subset of CM2M devices is scheduled during each
time slot, and the available spectrum is assigned among them without causing interference to
PUs.
3.2.2 Spectrum Aggregation Model
In a typical case of spectrum assignment, each channel is formed from a continuous spectrum
fragment; thus, it is not viable to use small spectrum fragments which are less than the user’s
bandwidth requirements. For example, consider a CM2M network where every machine needs
4 MHz channel bandwidth, and the available spectrum consists of two spectrum fragments of
4 MHz, and four spectrum fragments of 2 MHz (Fig. 3). For continuous spectrum allocation,
the 2 MHz spectrum fragments cannot be used by any machine. Thus, a continuous spectrum
assignment mode can only back two devices for communication (2 x 4 MHz). However, a
spectrum aggregation-enabled device can exploit fragmented segments of the spectrum by
utilizing specialized air interface techniques, such as DOFDM. In Figure 3.3, if a number of
small spectrum fragments are aggregated into a wider channel, then 16 MHz of unused
spectrum is available to support four CM2M devices (4 x 4 MHz).
Figure 3.3: Aggregation of Disjointed Spectrum Fragments
Because of the limited aggregation ability of the RF frontend, only those channels can be
aggregated that reside within a range of MAS. With this threshold, some spectrum fragments
Page 3 - 8
may not be aggregated because their span is larger than the MAS. The developed algorithm
considers MAS, for the sake of simplicity, the following assumptions are taken into
consideration:
1. All CM2M devices have the same aggregation capability (e.g., MAS for all devices is the
same).
2. The guard band between adjacent channels is neglected.
3. The bandwidth requirement of each device and bandwidth of each channel are an integer
multiple of sub-channel bandwidth Δ𝑓; which is the smallest unit of bandwidth (in fact, the
smaller fragments would demand excessive filtering to limit adjacent channel interference),
for example:
𝑟𝑛 = 𝑤𝑛 . Δ𝑓, 𝑤𝑛 ∈ ₦, 1 ≤ 𝑛 ≤ 𝑁 (3.1)
𝐵𝑊𝑚 = 𝑘𝑚 . Δ𝑓, 𝑘𝑚 ∈ ₦, 1 ≤ 𝑚 ≤ 𝑀 (3.2)
where ₦ is the set of natural numbers, 𝑤𝑛 is the number of requested sub-channels by CM2M
ɸn, 𝑘𝑚 is the number of sub-channels in 𝑦𝑚, and 𝐵𝑊𝑚 is the bandwidth of 𝑦𝑚.
The total available spectrum (e.g., M channels) is subdivided into multiple sub-channels. If the
available spectrum band consists of 𝐴 sub-channels, (e.g. total available bandwidth is 𝐴. 𝛥𝑓)
then:
𝑦𝑚 = ⋃ 𝑦��, 𝑚
𝑘𝑚
𝑖=1
. 𝑘𝑚 =𝐵𝑊𝑚
𝛥𝑓 , where 1 ≤ 𝑚 ≤ 𝑀 (3.3)
𝑨𝑨 = ∑ 𝑘𝑚
𝑀
𝑚=1
(3.4)
Page 3 - 9
Where 𝑦𝑚 𝑎𝑛𝑑 𝑘𝑚 sub-channels and ��𝑖, 𝑚 represent the 𝑖𝑡ℎ sub-channel of 𝑦𝑚. Each
��𝑖, 𝑚 can be represented in an interval defined as [ℱ𝑖,𝑚 𝐿 , ℱ𝑖,𝑚
𝐻 ]; where the ℱ𝑖,𝑚,𝐿 , and ℱ𝑖,𝑚
𝐻 ,
are the lowest and highest frequency of ��𝑖, 𝑚.
ℱ𝑖,𝑚 𝐻 − ℱ𝑖,𝑚,
𝐿 = 𝛥𝑓, For, 1 ≤ 𝑖 ≤ 𝑘𝑚 and 1 ≤ 𝑚 ≤ 𝑀 (3.5)
Based on this new sub-channel indexing, matrices LL and AA can be rewritten as:
𝑳𝑳∗= { 𝐿𝑛,𝑎∗ , 𝐿𝑛,𝑎
∗ = 𝐿𝑛,𝑚}𝑁𝑥 𝐴
(3.6)
𝑨𝑨∗= { 𝐴𝑛,𝑘,𝑎∗ , 𝐴𝑛,𝑘,𝑎
∗ = 𝐴𝑛,𝑘,.𝑚}𝑁𝑥𝑁𝑥𝐴
(3.7)
If
1 ≤ 𝑎 ≤ 𝑘1 𝑓𝑜𝑟 𝑚 = 1
and
∑ 𝑘𝑗
𝑚−1
𝑗=1
< 𝑎 ≤ ∑ 𝑘𝑗
𝑚
𝑗=1
𝑓𝑜𝑟 1 ≤ 𝑚 ≤ 𝑀
where 𝑎 represents the index of each sub-channel within the available spectrum. The sub-
channel assignment matrix
DD = {𝑑𝑛,𝑐 |𝑑𝑛,𝑐 ∈ {0,1}}𝑁𝑥𝐴
Page 3 - 10
is the 𝑁 𝑥 𝐴 binary matrix representing sub-channels allocated to CM2M devices for
aggregation such that 𝑑𝑛,𝑎=1, if and only if sub-channel 𝑎 is available to ɸn and 0 otherwise.
𝐁 = {𝑏𝑛 = 𝛥𝑓. ∑ 𝑑𝑛,𝑎𝐴𝑎=1 }
𝑁𝑥 1
B defines as the reward vector represent the total bandwidth that is assigned to each CM2M
device during scheduling time for a given sub-channel assignment.
3.3 Problem Formulation
3.3.1 Optimisation problem
One of the key objectives of the deployment of the CM2M network is to boost spectrum
utilization. To accomplish/attain etc. this crucial goal, the model optimizes network utilization
to maximize the total bandwidth that is assigned to CM2M devices and the
technique/methodology/scheme, etc. is referred to as Maximising Sum of Reward (MSR):
MSR = ∑ 𝑏𝑛
𝑁
𝑛=1
(3.8)
To maximize MSR, the spectrum aggregation problem can be defined as a constrained
optimization problem as shown in (3.9).
𝑚𝑎𝑥𝑑 = ∑ 𝑏𝑛
𝑁
𝑛=1
(3.9)
subject to,
Page 3 - 11
𝑏𝑛 = 𝛥𝑓. ∑ 𝑑𝑛,𝑐
𝐴
𝑎=1
= {0 𝑖𝑓 ɸn 𝑖𝑠 𝑟𝑒𝑗𝑒𝑐𝑡𝑒𝑑,
𝑟𝑛 𝑖𝑓 ɸn 𝑖𝑠 𝑎𝑐𝑐𝑒𝑝𝑡𝑒𝑑, 𝑓𝑜𝑟 1 ≤ 𝑛 ≤ 𝑁 (3.10)
ℱ𝑑,𝑡 𝐻 − ℱ𝑒,𝑓
𝐿 ≤ 𝑀𝐴𝑆 (3.11)
𝑑𝑛,𝑎 = 0 if 𝐿𝑛,𝑎∗ = 0 𝑓𝑜𝑟 1 ≤ 𝑛 ≤ 𝑁 𝑎𝑛𝑑 1 ≤ 𝑎 ≤ 𝐴 (3.12)
𝑑𝑛,𝑎. 𝑑𝑘,𝑎 = 0 if 𝐴𝑛,𝑘,𝑎∗ = 1 for 1≤ 𝑛, 𝑘 ≤ 𝑁 𝑎𝑛𝑑 1 ≤ 𝑎 ≤ 𝐴 (3.13)
Expression (3.10) assures that rewarded bandwidth 𝑏𝑛 to each accepted ɸn must be equal to
the ɸn’s bandwidth demand 𝑟𝑛; if the CM2M network cannot satisfy the ɸn ’s bandwidth
request, ɸn is rejected and 𝑏𝑛 = 0.
If ℱ𝑒,𝑓 𝐿 ( 1 ≤ 𝑒 ≤ 𝑘𝑓 and 1 ≤ 𝑓 ≤ 𝑀) is the lowest frequency of an initial aggregated sub-
channel and ℱ𝑑,𝑡 𝐻 (1 ≤ 𝑑 ≤ 𝑘𝑡 and 1 ≤ 𝑡 ≤ 𝑀) is the highest frequency of a terminative sub-
channel, (3.11) guarantees that the range of allocated spectrum is equal to or less than MAS.
The DD must satisfy the interference constraints (3.12) and (3.13); expression (3.12) and (3.13)
guarantee that there is no harmful interference to PUs and other CM2M devices respectively.
3.3.2 Spectrum Aggregation Algorithm Based on Genetic Algorithm (GA)
Typically, the spectrum assignment problem has been classified as an NP-hard problem [10].
Herein, GA is employed to overcome the aggregation-based spectrum assignment problem to
gain faster convergence. GA is a stochastic search method that mimics the process of natural
evolution. In addition, it is easy to encode solutions for the spectrum assignment problem to
chromosomes in GA and compare the fitness value of each solution. The specific operations of
Page 3 - 12
the proposed algorithm, referred to as MSR Algorithm (MSRA), can be described in the
following steps (Figure 3.4).
Figure 3.4: MSR Algorithm Flow Chart
Encoding
Initialisation
Selection
Genetic operators
Termination
Stop criteria of
GA satisfied
Yes
No
Page 3 - 13
1. Encoding: In MSRA, a chromosome represents a possible conflict-free sub-channel
assignment. To decrease search space (by reducing redundancy in the data) and gain faster
solutions, a comparable scheme as described in [10] is adopted in the model. The model
also applies a mapping process between DD and the chromosomes based on the
characteristics of 𝑳𝑳∗ and 𝑨𝑨∗.
Only those elements of DD whose corresponding elements in 𝑳𝑳∗ take the value of 1 are
encoded, for example 𝑑𝑛,𝑎 = 0 , where (n, a) satisfies 𝐿𝑛,𝑎∗ = 0. As a result of this mapping,
the chromosome length is equal to the number of non-zero elements of 𝑳𝑳∗ and the search
space is greatly decreased. Based on a given 𝑳𝑳∗ length of the chromosome can be
calculated as:
∑ ∑ 𝐿𝑖,𝑗∗
𝐴
𝑗=1
𝑁
𝑖=1
2. Initialisation: During the initialization process, the initial population is randomly
generated based on a binary coding mechanism as applied in [81]. The size of the
population depends on |𝝓|and |Y|; for larger |𝝓|and |Y| population size should be
increased; where |. | indicates cardinality of a set.
3. Selection: The fitness value of each individual of the current population according to the
MSRA criteria defined in equations (3.8)-(3.10) is computed. According to the individual’s
fitness value, excellent individuals are selected and remain in the next generation. The
chromosomes with the highest fitness values then replace the ones with the lowest fitness
values via the selection process.
4. Genetic operators: To maintain the high fitness values of all chromosomes in a successive
population, crossover and mutation operators are applied. Two randomly selected
chromosomes are chosen in each iteration as the parents, and the crossover of the parent
chromosomes is carried out at the probability of crossover rate. In addition to selection and
crossover operations, a mutation at a certain mutation rate is performed to maintain genetic
diversity.
5. Termination: The stop criteria of GA are checked in each iteration. If they cannot be
satisfied, step three and four above are repeated. The number of maximum iteration and the
Page 3 - 14
difference of fitness value are used as the criteria to determine the termination of GA. The
population of chromosomes generated by initialization, selection, crossover, and mutation
may not satisfy the given constraints defined in (3.10)-(3.13).
To find feasible chromosomes that achieve all thresholds, a threshold-free process is employed
that has the following steps (in order):
1. Bandwidth requirements: The vector B as given in (3.3) is calculated. 𝑏𝑛 should either
equal 𝑟𝑛 or zero, otherwise all genomes are related to ɸn changed to zero.
2. MAS: To satisfy the hardware limitations of the transceiver, expression (3.11) should
be satisfied; otherwise all genomes related to ɸn are changed to zero.
3. No interference with PUs: expression (3.12) guarantees that CM2M transmissions do
not interfere with LUs transmissions; ensuring that the CM2M network does not harm
the PUs performance. If Expression (3.12) is not satisfied, all genomes related to ɸn
are hanged to zero.
4. CCI: Expression (3.13) guarantees that there is no harmful interference to other CM2M
devices. If expression (3.13) is not satisfied, one of two conflicted devices is chosen
at random, and then all genomes of the selected device are changed to zero.
To obtain higher spectrum utilization and faster convergence, after each generation, MSRA
assigns all unassigned spectrum to remaining CM2M devices randomly, whenever possible. At
the same time, MSRA assures all the thresholds defined in (3.10)-(3.13) are satisfied at all
times.
3.4 Simulation Results
In this section, a set of system-level performance results are presented to compare and show
the efficiency of MSRA over MSA [14], AASAA [15] and Random Channel Assignment
Algorithm (RCAA) [16]. The simulation results demonstrate the high potential of the proposed
method in terms of spectrum utilization and system capacity. To assess the performance of the
network, independent of each device’s traffic distribution model, a backlogged traffic model
(known as a full-buffer model) is used where the packet queue length of every device is much
longer than can be scheduled during each scheduling time slot.
Page 3 - 15
Due to the random nature of the channel bandwidth and the devices bandwidth demand, Monte
Carlo simulations are performed, and each simulation scenario is repeated 100,00 times. The
default parameters used in the simulations are listed in Table 3.1, where U(1; 20) represents
the discrete uniform random integer numbers between 1 and 20. Each of the channels is
modeled as a flat Rayleigh channel with path loss model of 𝑃𝐿 = 128: 1 + 37: 6 𝑙𝑜𝑔10 𝑅
(R is in km) and penetration loss of 20 𝑑𝐵. The mean and standard deviation of log-normal
fading are zero and 8 𝑑𝐵 respectively.
In the simulation model, the CM2M devices are located randomly without restrictions within
a rectangular area of 2 𝑘𝑚 𝑋 1 𝑘𝑚. All channels are randomly selected between 54 MHz and
806 MHz television frequencies (Channels 2-69), to investigate the simulation results
effectively, the following terms are defined and used in the model analysis [14, 15]:
Table 3.1: Simulation parameters
Parameter Value
𝛥𝑓 The Bandwidth of sub-channel 1 MHz
MAS MAS 40 MHz
BWm The bandwidth of 𝑦𝑚 Δ . U (1; 20)
rn Bandwidth demand of ɸn Δ . U(1; 20)
Total Transmit Power 26 dBm (400 mW)
Scheduling Time Slot 1 ms
Traffic Model Backlogged
Backlogged
Population Size 20
Number of Generations
10
Mutation Rate
0:01
Crossover Rate 0:8
Page 3 - 16
1. Spectrum Utilisation: referred to as 𝑢 which is defined as the ratio of the sum of rewarded
bandwidth to the sum of all available bandwidth, e.g.
𝑢 = ∑ 𝑏𝑛
𝑛𝑁=1
∑ 𝐵𝑊𝑚𝑀𝑚=1
(3.14)
2. Network Load: referred to as Ƚ which is defined as the ratio of the sum of all CM2M
devices bandwidth requirements to the sum of all available bandwidth, e.g.
Ƚ = ∑ 𝑟𝑛
𝑛𝑁=1
∑ 𝐵𝑊𝑚𝑀𝑚=1
(3.15)
3. Number of Rejected Devices: rejected devices are those machines that are not assigned
any spectrum in a certain scheduling time slot.
3.4.1 Scenario-I: Without Co-Channel Interference
In this scenario, the performance of MSRA is compared with the SOTA algorithms including
MSA [14], AASAA [15], and RCAA [16], when the CCI among CM2M devices is not
considered. Therefore, the system model assumes that CM2M devices transmissions do not
overlap with the transmission of other CM2M devices using the same channel. For M = 30, Ƚ
increases by increasing the number of CM2M devices from 5 to 60. Figure 3.5 shows that when
the number of CM2M devices increases, spectrum utilization also increases in all three
methods, but MSRA utilizes all available whitespace in various network loading conditions
more efficiently than MSA, AASAA, and RCAA.
This can be explained by the fact that in the case of higher Ƚ, the network can allocate better
segments of the spectrum to users because of higher multi-user diversity. In addition, due to
using a stochastic search method, MSRA achieves a near to optimum solution in comparison
with other SOTA solutions which are based on approximate algorithms. For MSRA, when Ƚ is
higher than three, the CM2M network becomes saturated due to the lack of available spectrum.
However, for the remaining methods, there are still unassigned spectrum slices.
Page 3 - 17
Figure 3.5: The Impact of Varying Network Load Conditions on Spectrum Utilisation
(Scenario-I: Without CCI)
3.4.2 Scenario-II: With Co-Channel Interference
In this scenario, CCI exists among CM2M devices and MSRA compared with AASAA and
RCAA. As MSA does not inherently consider CCI, the model here does not include MSA for
comparison. Figure 3.6 shows spectrum utilization, according to different network loads by
increasing the number of CM2M devices from 5 to 55 when there are only seven available
channels (e.g., M = 7). As shown in Figure 3.6, MSRA outperforms AASAA and RCAA for
different network loads. Similar to when compared to Scenario-I, MSRA utilizes TVWS even
better, because some CM2M devices in the network may reuse spectrum previously used by
other devices in the CM2M network.
Page 3 - 18
Figure 3.6: The Impact of Varying Network Load Conditions on Spectrum Utilisation
(Scenario-II: With CCI)
Figure 3.7 represents the number of rejected CM2M devices when the network load increases.
The number of rejected CM2M devices increases with the network load; MSRA has a lower
number of rejected CM2M devices (or more satisfied devices) than AASAA and RCAA for
different network loads. MSRA optimizes spectrum utilization by admitting devices with better
channel quality to the network and allocates spectrum resources more effectively. Figure 3.7
implies that MSRA increases the capacity of the network, which is vital for M2M networks
due to the huge number of devices.
Page 3 - 19
Figure 3.7: The Impact of Varying Network Load Conditions on the Number of Rejected
CM2M Devices (Scenario-II: With CCI).
3.4.3 Convergence of MSRA
Because of the nature of genetic programming [17][12], it is arguably impossible to make
formal guarantees about the number of fitness evaluations needed for an algorithm to find an
optimal solution. However, herein, computer experiments are performed to show the impact of
a number of generations on the performance of MSRA. The system parameters used in the
section for simulation are listed in Table 3.2. For convergence studies, N is set/chosen, etc to
be 200 and M = 10. Figure 3.8 shows the best fitness value (MSRA) for a population in a
different number of generations. As shown in Figure 3.8, the performance of the algorithm is
enhanced as the number of generations increases; after roughly 34 generations, the fitness value
saturates at an optimal value which shows the effectiveness of using GA for spectrum
assignment using spectrum aggregation.
Meanwhile, Figure 3.9 illustrates the distribution of processing time for MSRA to find an
optimal solution. As shown in Figure 3.8, 85% of the time, MSRA finds an optimum solution
in less than the scheduling time slot (1 ms) and 15 % of the time it takes more than the
Page 3 - 20
scheduling time slot which is a good rate comparing with the AASAA and RCAA which
usually takes more than (3 ms) to find the optimal solution.
Figure 3.8: The Impact of Number of Generations on MSRA Results
Table 3.2: System Parameters
Parameter Value
M 10
N 200
Processor Intel Core i7-3667U 2.00 GHz
Memory (RAM) 4 GB
OS Windows 7 (64-bit)
Page 3 - 21
Figure 3.9: Distribution of Processing Time for MSRA to find an Optimal Solution
Furthermore, Lobo in [12] provided a theoretical and empirical analysis of the time complexity
of traditional simple GAs. According to [12], GA has time complexities of
Ϭ (∑ ∑ 𝐿𝑖,𝑗∗
𝐴
𝑗=1
𝑁
𝑖=1
)
which is dependent on the length of each chromosome. The linear time complexity for GA
occurs because the population size grows with the square root of chromosome length. In this
problem, the complexity depends on chromosome length which itself is dependent on a number
of users and radio resources. The time complexity presented herein is for the worst-case
scenario (Ϭ) when the population size is assumed fixed at the maximum number of generations.
Page 3 - 22
3.5 Summary
This chapter introduces an aggregation-aware spectrum assignment algorithm using a genetic
algorithm. The proposed algorithm maximizes spectrum utilization for CM2M devices as a
criterion to realize spectrum assignment. Moreover, the introduced algorithm takes into account
the realistic thresholds of co-channel interference and increase aggregation span. The
performance of the proposed algorithm is validated by simulations, and the results are
compared with algorithms available in the current literature (e,g, RCA, MSA, and MSRA). The
proposed algorithm reduces the number of rejected devices and enhances spectrum utilization
of the CM2M network. The algorithm increases the capacity of the network which is vital for
M2M networks.
Page 3 - 23
References for Chapter 3
[1] Cisco visual networking index: Global mobile data traffic forecast update 2014-2019
white paper. [Online, 2017]. Available: http://www.cisco.com/c/en/us/solutions/
collateral/serviceprovider/visual-networking/indexvni/whitepaperc11-520862.html.
[2] S. Rostami, K. Arshad, and K. Moessner, “Order-statistic based spectrum sensing for
cognitive radio,” Communications Letters, IEEE, vol. 16, no. 5, pp. 592–595, 2012.
[3] Quoc Duy Vo, Joo-Pyoung Choi “Green Perspective Cognitive Radio-based M2M
Communications for Smart Meters”, Tutorials, IEEE 978-1-4244-98072010.
[4] Y. Zhang, R. Yu, M. Nekovee, Y. Liu, S. Xie, and S. Gjessing, “Cognitive machine-
to-machine communications: visions and potentials for the smart grid,” Network,
IEEE, vol. 26, no. 3, pp. 6–13, May 2012.
[5] Federal Communications Commission, [Accessed 29/06/2017]: https://www.fcc.gov/.
[6] M. Wylie-Green, “Dynamic spectrum sensing by multiband OFDM radio for
interference mitigation,” in New Frontiers in Dynamic Spectrum Access Networks,
2005. DySPAN 2005. 2005 First IEEE International Symposium on, Nov 2005, pp.
619–625.
[7] J. Poston and W. Horne, “Discontiguous OFDM considerations for dynamic spectrum
access in idle TV channels,” in New Frontiers in Dynamic Spectrum Access
Networks, 2005. DySPAN 2005. 2005 First IEEE International Symposium on, 2005,
pp. 607–610.
[8] R. Rajbanshi, A. M. Wyglinski, and G. J. Minden, “An efficient implementation of
NC-OFDM transceivers for cognitive radios,” in 2006 1st International Conference
on Cognitive Radio Oriented Wireless Networks and Communications, June 2006, pp.
1–5.
[9] 3GPP, carrier aggregation for LTE. [Online]. Available: http:
//www.3gpp.org/ftp/information/workn plan/description releases.
[10] Z. Zhao, Z. Peng, S. Zheng, and J. Shang, “Cognitive radio spectrum allocation using
evolutionary algorithms,” Wireless Communications, IEEE Transactions on, vol. 8,
no. 9, pp. 4421–4425, 2009.
Page 3 - 24
[11] K. Arshad, M. Imran, and K. Moessner, “Collaborative spectrum sensing optimization
algorithms for cognitive radio networks,” International Journal of Digital Multimedia
Broadcasting, vol. 2010, no. 1, pp. 1–20, 2010.
[12] F. G. Lobo, D. E. Goldberg, and M. Pelikan, “Time complexity of genetic algorithms
on exponentially scaled problems,” in Proceedings of the genetic and evolutionary
computation conference. Morgan-Kaufmann, 2000, pp. 151–158.
[13] Y. Li, L. Zhao, C. Wang, A. Daneshmand, and Q. Hu, “Aggregation based spectrum
allocation algorithm in cognitive radio networks,” in Network Operations and
Management Symposium (NOMS), 2012 IEEE, 2012, pp. 506–509.
[14] F. Huang, W. Wang, H. Luo, G. Yu, and Z. Zhang, “Prediction-based Spectrum
aggregation with hardware limitation in cognitive radio networks,” in Proceedings of
the IEEE 71st Vehicular Technology Conference (VTC '10), pp. 1–5, May 2010.
[15] D. Chen, Q. Zhang, and. Jia, “Aggregation aware spectrum assignment in ad-hoc
cognitive networks,” in Cognitive Radio Oriented Wireless Networks and
Communications, 2008. CrownCom 2008. 3rd International Conference on, 2008, pp.
1–6.
[16] E. Anifantis, V. Karyotis, and S. Papavassiliou, "A Markov Random Field framework
for channel assignment in Cognitive Radio networks," 2012 IEEE International
Conference on Pervasive Computing and Communications Workshops, Lugano,
2012, pp. 770-775. doi: 10.1109/PerComW.2012.6197617.
[17] Fang Ye, Rui Yang, and Yibing Li. Genetic algorithm based spectrum assignment
model in cognitive radio networks. In Information Engineering and Computer Science
(ICIECS), 2010 2nd International Conference on, pages 14, Dec 2010.
Page 4 - 1
CHAPTER 4
ENERGY EFFICIENT SPECTRUM
SENSING AND SHARING
Page 4 - 2
4.1 Introduction
M2M communication is the future of smart things; as discussed in the previous Chapters,
hundreds of millions of devices will be connected in the near future for a number of different
applications. M2M communication is very important due to its cost reduction, flexibility, time
efficiency, and accuracy. M2M communications will face many challenges due to a high
number of devices, cost, and the quality and reliability of the providing services [1, 2]. These
challenges are mainly related to energy consumption, spectrum limitations, data rate, and
security.
There are currently many attempts being made to address M2M communications challenges
such as energy and spectrum efficiency. However, little consideration has been given to the
idea of using cognitive radio as a solution. Cognitive radio is an intelligent technology which
can be designed and configured to cope with the M2M communications problems.
As mentioned in Chapter 2 a number of studies have considered cognitive radio technology for
better energy efficiency, one of these addressed spectrum discovery schemes such as the non-
cooperative, cooperative and time-division energy efficient schemes [3] for improved energy
efficiency in CM2M communications. The authors in [4] addressed optimal power allocation
to improve quality of service and energy efficiency in a CM2M network. In [5] the energy
efficiency in CR wireless devices is addressed. The distributed sensing approach optimizes
power efficiency with thresholds on the minimum desired detection probability and the
increase of the permissible probability of false alarms by selecting the sensing and sleeping
design parameters.
Little efforts have considered spectrum handoffs and the wait/switch trade-off in a CM2M
network with multiple CM2M devices (e.g., CM2M gateways). In addition, no previous work
has considered the scenario of a number of CM2M gateways working together and addressed
the collision probability among them.
This Chapter proposes an energy efficient mechanism for the CM2M network by optimizing
the spectrum sensing and switching schemes. The proposed mechanism guarantees to sense
reliability and users’ throughput constraints simultaneously. To optimise the total energy cost,
the mechanism ensures that spectrum handoff is not used excessively. Instead,
Page 4 - 3
CM2M gateways may occasionally decide to stop sending data and remain in a sleep state on
their current channel for a specific duration of time.
The rest of this Chapter is organised as follows: section (4.2) presents the system model, section
(4.3) addresses the problem and presents the solution, section (4.4) shows the optimality of our
mechanism through simulation and discussion, and Section (4.5) concludes the chapter.
4.2 Spectrum Sensing and Sharing System Model
The system model considers working with Q number of secondary transmission CM2M
gateways and primary transmission (Figure 4.1). As shown in Figure 4.1 CM2M gateways
connected M2M device domain and network domain together and the the final signal can be
send to the backend (application domain). Furthermore, the system considers the following
assumptions:
1. ᶆ random channels shared between secondary users and primary users, and Ɲ is the
number of frames.
2. The secondary transmission (CM2M gateways) is slotted in via periodic sensing at
specific periods.
3. Each frame builds a transmission slot of period T and from sensing slot of durations 𝜏𝑠
a transmission slot of period T, where only one of the ᶆ channels is assigned to the
CM2M gateways.
4. The CM2M gateways at the start of each transmission slot may prefer to send
information/data on the sensed channel or select another available channel.
5. The PU transmission is presumed to be continuous and follow an off-on- traffic pattern
for each channel [6].
6. The primary user transmission probability (ON/OFF) presumed to be the same for all
the channels.
7. The average received SNR of the primary transmission signal for all channels is
assumed to be the same through one packet of data transmission.
Page 4 - 4
Figure 4.1: CM2M System with a Number of Secondary Transmission CM2M Gateways
Figure 4.2 shows the CM2M gateways performance when sending a packet of information/data
assuming energy detection is applied to jointly sense the channels while the CM2M gateways
apply wideband sensing to check the other free channels. Furthermore, Figure 4.2 shows how
CM2M gateways sometimes choose to sleep (power off) when the current channels busy; and
the collision between the CM2M gateways if they decided to jump at the same time into the
same channel.
The false alarm probability and the detection probability, are defined as 𝑃𝑓 and 𝑃𝑑. The
received signal is sampled at sampling frequency 𝑓𝑠, bandwidth 𝐵𝑊 and the frequency band
with carrier frequency 𝑓𝑐. The system model presumed when the primary transmission is
active, the received signal at the CM2M gateways under assumption 𝑆0 and can be represented
as :-
𝑎(𝑛) = 𝑠(𝑛) + 𝑢(𝑛)
Page 4 - 5
Figure 4.2: CM2M Gateways Performance in Ɲ Packet of Data
Where the system considers the following assumptions [7]:
• The first assumption is that the noise u(n) is Gaussian, identically and independently
distributed in a random way with mean zero and; E|u(n)|2 = σu2;
• The primary signal 𝑠(𝑛) is assumed to be a real-valued Gaussian signal [7].
When the primary user under idle state, the output under assumption 𝑆1. The received signal
is given by
𝑎(𝑛) = 𝑢(𝑛)
Two reliable probabilities are considered for spectrum sensing. The first is the probability of
false alarm, which is determined under assumption stated in S0, is the probability where the
algorithm falsely declares the existence of primary transmissions. The second is the probability
of detection, which is determined under the assumption stated in 𝑆1, is the probability where
the algorithm correctly detects the existence of primary transmissions [8].
For the primary user’s transmission side to gain better protection, a higher probability of
detection should be guaranteed. From the CM2M gateways transmission side, a lower
Page 4 - 6
probability of false alarm should be guaranteed for better CM2M gateways transmission (QoS).
Therefore, for an efficient algorithm, the probability of false alarm should be the minimum
possible, and the probability of detection should be the maximum possible. The probability of
false alarm related to the target false alarm probability ��𝑓 and the target probability of
detection ��𝑑 for similar system are given in [8] and as follows:
𝑃𝑓 = 𝒬(√2γ + 1 𝒬−1 (��𝑑) + √𝜏 𝑓𝑠γ) (4.1)
where γ is the SNR regime, 𝑓𝑠 is the sampling frequency, 𝜏 is the CM2M sensing time.
Similarly, the probability of detection related to the target false alarm probability ��𝑓 and the
target probability of detection ��𝑑 for similar system are given in [8] and as follows
𝑃𝑑 = 𝒬 (1
√2γ+1 𝒬−1 (��𝑓) + √𝜏 𝑓𝑠γ) (4.2)
Next, the system considers ��𝑑 and ��𝑓 are the same for all channels. From [88] it is known,
for a stable sampling frequency 𝑓𝑠, there is a minimum sensing period such that the reliability
constraints are achieved (false alarm and detection probabilities). The minimum required
sensing time is defined by 𝜏𝑠𝑚𝑖𝑛 .
𝜏𝑠𝑚𝑖𝑛 =
1
γ𝑓𝑠(𝒬−1(��𝑓) − 𝒬−1(��𝑑)√2γ + 1 )2 (4.3)
To decrease the energy consumption of the CM2M gateways due to spectrum sensing, 𝜏𝑠𝑚𝑖𝑛
sensing duration should be satisfied. The total energy consumption should be achieved with
optimal 𝜏𝑠 to achieve higher energy efficiency in the model [6 , 9].
The CM2M gateways will choose before sending data whether to select another empty channel
or remain on the current sensed channel after collecting credible knowledge of the availability
Page 4 - 7
of other surrounding channels. The system model assumed to deter CM2M gateways from
switching and selecting another channel; they have to apply transmission for period T until the
next sensing slot arrives (Figure 4.2.) If CM2M gateways remain on their current channels,
CM2M gateways must decide whether to simply refrain from sending data until the next
sensing slots become available, or to proceed with data transmission for a period of time T
using other channels.
The delay due to spectrum handoff and the energy cost during the transmission slot is assumed
to be negligible in the case where the CM2M gateways devices stay on the current channels
with power off. However, the designed model switch/select-wait/sleep considers the trade-off
between the activities of the CM2M gateways in cases of energy savings and throughput [10,
11]
For instance, one of the CM2M gateways may stay and sleep on the current channel and
continue sending data when the current channel is sensed as free, and the other channels sensed
as busy. Because neither energy savings nor throughput will improve if the CM2M gateways
switch to another channel. Therefore, the CM2M gateways must remain on the current channel
and sleep for a period of T seconds when all ᶆ channels busy, as energy costs will increase
when the CM2M gateways decide to switch and send data on the other channels without
enhancing the data rate of the CM2M gateways [8, 6].
If the CM2M gateways sense the current channel is busy and one of ᶆ channels is sensed as
unoccupied by the primary user, they need to choose whether to spend energy to switch and
select a free channel so that the CM2M gateways may be able to procced and make data
transmission or to stay on the current channel and sleep to spare energy at the cost of reduced
throughput [12]. In such a case, CM2M gateways are assumed to switch to another free channel
with a probability of (1 – 𝐿𝑠) or stop transmission and wait on the current channels with a
probability of 𝐿𝑠 [13].
Page 4 - 8
4.3 System Problem and Solution
4.3.1 Problem
The problem is formulated by jointly looking for both optimal values of ( 𝜏𝑠, and 𝐿𝑠).
subsequently, the energy efficiency is maximized with the satisfied throughput. The
optimization problem can be formulated as follows:
𝑚𝑖𝑛 𝜏𝑠 𝐿𝑠 𝑄𝑋(𝜏𝑠, 𝐿𝑠)
subject to
𝑃𝑑(𝜏𝑠) ≥ ��𝑑
𝑃𝑓 (𝜏𝑠) ≤ ��𝑓
𝑄ℛ(𝜏𝑠, 𝐿𝑠) ≥ 𝑟 (4.4)
Where X is the overall average energy cost needed to finish sending a single packet of
information/data, 𝑆 is the time needed for the CM2M gateways to send a packet of data, ℛ is
the average data rate or throughput, and 𝑟 is the minimum throughput that should be achieved
in the system, respectively.
4.3.2 Problem Solution
The system model defines 𝑃3 as the switching probability to another idle channel and apply
transmission of T for a number of CM2M gateways and one primary user link as given as
follows:
𝑃3(𝜏𝑠, 𝐿𝑠) = (1 − 𝑃1)(1 − 𝑃𝑏)(1 − 𝐿𝑠) (4.5)
Where 𝑃1 is the probability of a channel being free, and 𝑃𝑏 is the probability of a channel being
busy. The system model is assumed to work with a number of CM2M gateways thus, the
Page 4 - 9
probability of collision between CM2M devices need to be manged in case they decided to
switch to the same idle channel at the same time. For this, the system model considered the
transmissions of CM2M gateways for such a case equal to zero. At the same time, the system
model assumed 𝑝𝑓𝑐 as the probability of collision between CM2M gateways. Hence, the new
probability of switching to another idle channel and performing transmission of T is given in
equation (4.6):
𝑃3(𝜏𝑠, 𝐿𝑠) = (1 − 𝑃1)(1 − 𝑃𝑏)(1 − 𝐿𝑠)(1 − 𝑝𝑓𝑐) (4.6)
Hence, the overall energy cost for such a system, including the energy cost due to spectrum
handoff, spectrum sensing, and data transmission, can be calculated by the equation below: -
𝑋(𝜏𝑠, 𝐿𝑠) = 𝑆𝐸𝑡 + Ɲ 𝜏𝑠𝐸𝑠 + Ɲ𝑃3 𝐽𝑠𝑤 (4.7)
where 𝐽𝑠𝑤 is defined as the energy cost for single channel switching, in the unit of joules,
where 𝐸𝑡 and 𝐸𝑠 is defined as energy consumption per second due to sending data and sensing
both in the unit of watts, presuming that 𝐸𝑡 , 𝐸𝑠 and 𝐽𝑠𝑤 are known as a given for CM2M
gateways [8]. Similar power consumption is presumed when the CM2M gateways switch from
one channel to another sensed as free. In such a situation, if there are free channels and the
secondary user tries to switch one will be picked at random without compromising the overall
average energy cost.
𝐶0 = 𝑙𝑜𝑔 (1 + 𝑆𝑁𝑅) bits/s/Hz is assumed to be the CM2M gateways throughput, the bits
that are successfully sent in one transmission duration of T assumed to be 𝐶0𝑇.
There are two scenarios where the CM2M gateways choose to remain on the current sensed
channel: when there is a minimum of one other free channel, but the CM2M gateways decide
not to switch with a probability of 𝐿𝑠 and when all M channels are sensed as occupied. The
average data rate of the CM2M gateways is defined by
Page 4 - 10
ℛ(𝜏𝑠,𝐿𝑠) = 1 −𝑃𝑐 + 𝑃3 (1 − 𝑃𝑒 )𝐵𝑡
𝜏𝑠 + T (4.8)
where 𝐵𝑡 is number of bits sent in one transmission slot, 𝑃𝑐 is the probability of channels truly
sensed as busy, 𝑃𝑒 those mistakenly sensed as busy [14], and 𝑟 assumed to be small due to the
CM2M system’s small data rate requirements. It is observed from the equation given in (4.8)
that, for a given 𝜏𝑠, ℛ is decreasing with 𝐿𝑠. This is true, because the longer the CM2M
gateways remain on the current channel without sending any data (with a larger 𝐿𝑠), the less
throughput can be achieved within a given T duration. In addition, it is known from (4.4) that
the optimal 𝐿𝑠 is measured by the last threshold (𝑟), once it can achieve that threshold it can be
absolutely fulfilled with the parameters given in the model. Accordingly, the optimal 𝐿𝑠 is
easily maximum the allowable of 𝐿𝑠 that fulfilled the last threshold in (4.4). Which means the
optimal 𝐿𝑠 that solves (4.4) is given by
𝐿𝑠𝑜𝑝𝑡
(𝜏𝑠) = 1 −(𝜏𝑠 + 𝑇)𝑟 − 𝑃𝑐1 𝐵𝑡
𝐵𝑡(1 − 𝑃1)(1 − 𝑃𝑒)(1 − 𝑃𝑏)(1 − 𝑝𝑓𝑐) (4.9)
where 𝐿𝑠𝑜𝑝𝑡
is obtained by letting ℛ(𝜏𝑠, 𝐿𝑠) = 𝑟, and from (4.6) and (4.8) it’s possible to get
the 𝐿𝑠𝑜𝑝𝑡
.
Furthermore, the first two thresholds 𝑃𝑑(𝜏𝑠) ≥ ��𝑑 and 𝑃𝑓 (𝜏𝑠) ≤ ��𝑓 in (4.4) functionally
designates that 𝜏𝑠 ≥ 𝜏𝑠𝑚𝑖𝑛, and from (4.3) its known that 𝜏𝑠
𝑚𝑖𝑛 can satisfy the first two
constraints. Furthermore, from (4.7) and (4.8) its known 𝜏𝑠𝑚𝑖𝑛 can achieve the minimum
energy consumption in the system with guaranteed ℛ > 𝑟. This can be explained by the
sensing duration of CM2M gateways: the less sensing time, the less energy required to sense
the channels, which leads ultimately to decreased energy consumption in the system overall.
Now, it’s clear that 𝜏𝑠𝑚𝑖𝑛 can satisfy the three constraints given in (4.4) and at the same time
achieve the minimum energy consumption in the system. Therefore, 𝜏𝑠𝑚𝑖𝑛 can be consider the
Page 4 - 11
𝜏𝑠𝑜𝑝𝑡𝑖𝑚𝑎𝑙 for the system that satisfies all constraints and reduces energy consumption in the
system.
4.4 Simulation and Discussion
The efficiency of the designed spectrum sensing and accessing mechanism was demonstrated
through Matlab simulations (Monte Carlo), and Table 4.1 shows the simulation setting. In
simulations, the primary user is presumed to be a Quadrature Phase Shift Keying (QPSK)
modulated signal with a bandwidth of 6MHz [8]. The sampling frequency is the same as the
bandwidth of the primary user.
In the simulation, the following assumptions were made; ᶆ is assumed to be 6 channels
available for spectrum sharing between the CM2M gateways and the primary users, the period
of each transmission slot is T = 0.8s, the period of a packet of information/data is 𝑆 = 8s, and
the energy needed for spectrum sensing and sending data are 𝐸𝑡 = 69.5 mW 𝑎𝑛𝑑 𝐸𝑠 = 40
mW, respectively [15]. Furthermore, the received SNR of the primary user signal is assumed
to be γ = -10 dB, and the average throughput of CM2M gateways is defined as 𝑅0 = (1 −
𝜌)𝐶0 . The throughput constraint is defined as 𝑟 = 𝜇𝑄𝑅0 where, the throughput coefficient
𝜇 ∈ [0, 0.6] and 𝑄 the CM2M gateways numbers.
Table 4.1: Value Setting of Simulations [8]
Parameter Value Description
ᶆ 6 Available channel number
𝑆 8 s Duration of one packet data
T 0.8s Duration of each transmission
𝛾 -10dB SNR
𝐽𝑠𝑤 40mJ
The energy consumption for single channel switching
𝝆 0.4 The probability of a channel being busy
��𝑑 0.9 The target probability of detection
��𝑓 0.1 The target probability of false alarm
𝑃𝑓𝑐
0.3 Collision probability of CM2M gateways
Q 2 CM2M gateways number
Page 4 - 12
The energy consumption of the CM2M gateways in Table 4.2 is energy cost as a function of
the throughput coefficient 𝜇 for 𝐽𝑠𝑤 = 40 𝑚𝐽. It is obvious that when the value of 𝐿𝑠
increases, better energy efficiency is achieved, and the reason is that when 𝐿𝑠 increases, fewer
CM2M gateways will switch to a new channel and more CM2M gateways numbers will stay
in the current channel and sleep for a period of time (T).
Accordingly, the optimal energy efficiency of X will be achieved when 𝜏𝑠 is equivalent to
𝜏𝑠𝑚𝑖𝑛 as explained in the aforementioned section. This is true because the lower the value of
𝜏𝑠, the less energy will be spent in channel sensing, and the larger the T time duration
transmission to send more data and increase the throughput. The numerical results in Table 4.
2 proved the analysis given in previous sections.
Table 4.2: Energy Consumption in CM2M Devices
In Table 4.3, the developed mechanism results are compared with the ones reported in [8]
which used only a sensing/throughput mechanism. By using simulations with the same value
settings, it is shown that the developed mechanism (wait/switch trade-offs and sensing
/throughput) is more energy efficient than the one given in [8].
𝜏𝑠𝑜𝑝𝑡𝑖𝑚𝑎𝑙 in [8] only optimises the sensing/ throughput trade-off and is not optimal in the case
of reducing the overall energy cost, where it leads to more power cost compared with the state
when 𝜏𝑠𝑜𝑝𝑡𝑖𝑚𝑎𝑙 of the developed mechanism is applied. Bear in mind that 𝜏𝑠
𝑜𝑝𝑡𝑖𝑚𝑎𝑙 is created
when both the wait/switch trade-offs and sensing/throughput are simultaneously considered.
Energy consumption (mW)
Throughput coefficient
𝝉𝒔 Optimal 𝑳𝒔
1235 [0,0.6]
32 ms 0
1125 [0,0.6]
32 ms 0.5
900 [0,0.6] 32 ms 𝐿𝑠𝑜𝑝𝑡
Page 4 - 13
Table 4.3: Energy cost when optimal 𝐿𝑠 and 𝜏𝑠 are employed compared with the one in [8]
which employed only 𝜏𝑠
4.5 Summary
In this Chapter, an energy efficient mechanism for CM2M communication has been proposed.
The proposed mechanism simultaneously considers the wait/switch trade-off in terms of
channel switching probability and the sensing /throughput trade-off in terms of the duration of
sensing time. The proposed mechanism addresses the probability of collision between CM2M
gateways by considering the transmission of devices to be zero when a collision occurs, and
guarantees that the given constraints in terms of throughput and sensing reliability are always
satisfied. Simulation results show the efficiency of performing spectrum handoff and the
optimality of the mechanism that guarantees desirable throughput and reduced energy
consumption in the system.
Energy consumption (mW)
Throughput
coefficient
𝝉𝒔 𝐎𝐩𝐭𝐢𝐦𝐚𝐥 𝑳𝒔
1435 (0,0.6) [8] 41 ms 0
1350 (0,0.6) [8] 41 ms 0.5
1100 (0,0.6) [8] 41 ms 𝐿𝑠𝑜𝑝𝑡
1235 [0,0.6] 32 ms 0
1125 [0,0.6] 32 ms 0.5
900 [0,0.6] 32 ms 𝐿𝑠𝑜𝑝𝑡
Page 4 - 14
References for Chapter 4
[1] Quoc Duy Vo, Joo-Pyoung Choi “Green Perspective Cognitive Radio-based M2M
Communications for Smart Meters”, Tutorials, IEEE 978-1-4244-98072010.
[2] T. Yilmaz, R. Foster, and Y. Hao, “Detecting vital signs with wearable wireless
sensors, ” Sensors, vol. 10, no. 12, pp. 10837–10862, Dec. 2010.
[3] Fadlullah, Z., Fouda, M. M., Kato, N., Takeuchi, A., Iwasaki, N., & Nozaki, Y.
(2011, April). Toward Intelligent Machine-To-Machine Communications in Smart
Grid. IEEE Communications Magazine, 49(4), 60–65.
[4] llanko, K., Naeem, M., Anpalagan, A., & Androutsos, D. (2011). Energy-Efficient
Frequency and Power Allocation for Cognitive Radios in Television Systems. IEEE
Systems Journal, 10(1), 313-324.
[5] Hu W., Dinh T.L., Corke P., Jha S. Outdoor sensor net design, and deployment:
Experiences from a Sugar Farm. IEEE Pervasive Computer. 2012;11:82–91.
[6] Wang, S., Wang, Y., Coon, J. P., & Doufexi, A. (2012). Energy-efficient spectrum
sensing and access for cognitive radio networks (pp. 13–17). Vehicular.
[7] M. Hatton, The Global M2M Market in 2013. London, U.K.: Machina Research
White Paper, Jan. 2013.
[8] Liang, Y. C., Zeng, Y., Peh, E. C., & Hoang, A. T. (2008). Sensing-throughput trade-
off for cognitive radio networks. IEEE Transactions on Wireless Communications,
7(4), 1326–1337. doi:10.1109/TWC.2008.060869.
[9] Su, H., & Zhang, X. (2010). Power-efficient periodic spectrum sensing for cognitive
MAC in dynamic spectrum access networks. Proc. IEEE WCNC (pp. 1–6).
[10] Li, X., Zhao, Q., Guan, X., & Tong, L. (2010). Sensing and communication trade-off
for cognitive access of continues-time Markov channels. Proc. IEEE WCNC (pp. 1–
6). doi:10.1109/WCNC.2010.5506649.
[11] Lu, R., Li, X., Liang, X., & Lin, X. (2011). GRS: The green, reliability, and security
of emerging machine to machine communications. IEEE Communications
Magazine, 49(4), 28–35. doi:10.1109/MCOM.2011.5741143.
Page 4 - 15
[12] Yao, J. (2013). Cognitive machine-to-machine communications: Visions and
potentials for the smart grid. IEEE Network, 26(3), 6–13.
doi:10.1109/MNET.6201210.
[13] ETSI, Machine to Machine Communications (M2M): Use cases of M2M applications
for eHealth, ETSI TR 102 732, 2011.
[14] Hoang, A. T., Liang, Y.-C., Wong, D. T. C., Zeng, Y., & Zhang, R. (2009).
Opportunistic spectrum access for energy-constrained cognitive radios. IEEE
Transactions on Wireless Communications, 8(3), 1206–1211.
doi:10.1109/TWC.2009.080763.
[15] Willian D. de Mattos; Paulo R. L. Gondim” M-Health Solutions Using 5G Networks
and M2M Communications” Pages: 24 - 29, DOI: 10.1109/MITP.2016.52 2016.
Page 5 - 1
CHAPTER 5
SELECTIVE ANTENNA SENSING
Page 5 - 2
5.1 Introduction
M2M applications in the medical sector have become more significant for private and public
entities, such as monitors for temperature, blood pressure and blood oxygen levels, and
electrocardiograms [1]. This leads to improved efficiency in e-healthcare systems for instance
by allowing early patient discharge from healthcare facilities and less time spent in recovery
and reducing overall costs accrued by the patient’s family and allocated government spending
[2, 3].
However, as mentioned in previous Chapters, the use of wireless M2M communications in e-
healthcare systems creates new and complex interference scenarios. This intricacy is generated
by medical devices which are normally sensitive to electromagnetic interference [4] caused by
wireless antennas. The interference can cause many problems for the medical machines (e.g.
waveform distortion, automatic shutdown, and automatic restart), which can be hazardous to
patients. Furthermore, M2M applications in healthcare systems need to be energy efficient and
able to work for a long period (e.g., 10 years) without battery replacement, while at the same
time the quality of service should be sufficient to ensure the reliability thresholds required in
e-healthcare applications. Cognitive radio can help with avoiding spectrum interference and
improve spectrum and energy efficiency in the e-healthcare system [5, 6].
Figure 5.1: CM2M Network in Healthcare System
Page 5 - 3
From the previous Chapters, its known a number of studies have proved that a “spectrum
hands-off” operation can increase the data rate of the secondary users (e.g. CM2M gateways)
when there is a variety of channels free for spectrum sharing [7], where the CM2M gateways
can switch to another free channel and proceed with sending data when the current used channel
is busy [8, 9].
This Chapter, propose an energy efficient spectrum management mechanism for the CM2M e-
healthcare system as shown in Figure 5.1. The mechanism work illustrated in Figure 5.1 which
connected the M2M devices with CM2M gateways to carry out the traffic to the backend
through network domain which can be any networks (e.g., TVWS Bands). The proposed
mechanism guarantees to sense reliability, throughput, delay and collision probability
thresholds simultaneously. The rest of this Chapter is organised as follows: Section (5.2)
defines the system model and explains spectrum sensing and access mechanisms. Section (5.3)
formulates the optimization problem, while section (5.4) proposes a solution. Discussions and
simulation results are illustrated in Section (5.5), and Section (5.6) summarises the Chapter.
5.2 Selective Antenna Sensing System Model
The system model supposes a CM2M network working in an e-healthcare environment with
two CM2M gateways and a PU. Consider that there are ᶆ channels shared between the CM2M
gateways and a primary user. Energy detection is assumed to sense the channels, while the
CM2M gateways apply wideband sensing to check the availability status of all channels. The
additive noise is a zero-mean CSCG process. The system model considers a low SNR regime
(γ), and we presume that the SU is slotted in via periodic sensing at a specific time, each frame
consisting of a sensing slot of durations 𝜏𝑠 and a transmission slot of period T, where one of
the ᶆ channels is assigned to the CM2M gateways [10, 11].
The spectrum handoff delay assumed to be 𝑘𝑠𝑙 is also added after the sensing slot in the CM2M
gateways frames if the CM2M gateways decide to carry out channel switching after considering
the sensing result and other thresholds.
Figure 5.2 shows CM2M gateways at the start of each transmission slot may decide to send
information/data on the current channel, switch to another channel with a switch delay of 𝑘𝑠𝑙,
or remain on the current channel without sending any data. Moreover, the Figure shows the
primary transmission activity which follows an off-on traffic pattern and is assumed to be
Page 5 - 4
continuous [12]. The average received signal-to-noise ratio (SNR) of the PU’s signal for all
channels assumed to be the same, also the false alarm probability 𝑃𝑓 and the detection
probability 𝑃𝑑 assumed to be the same during one packet of data transmission [12]. The
received signal is presumed to be sampled at sampling frequency 𝑓𝑠, the frequency band with
carrier frequency 𝑓𝑐 and bandwidth 𝐵𝑊.
Figure 5.2: CM2M Gateways Performance in Transmitting Ɲ Packet of Data
The goal is to decrease energy costs for the CM2M gateways to send a packet of data under the
probability of both false alarm and detection thresholds. As discussed in [12], we can jointly
optimize the sensing slot length 𝜏𝑠 and the stay probability on the current channel 𝐿𝑠 so that
the energy consumption of one complete data packet transmission is reduced as follows:
( 𝜏𝑠𝑂𝑝𝑡𝑖𝑚𝑎𝑙 , 𝐿𝑠
𝑂𝑝𝑡𝑖𝑚𝑎𝑙) = min 𝜏𝑠 ,𝐿𝑠
𝑋 (𝜏𝑠, 𝐿𝑠) (5.1)
Page 5 - 5
where 𝐿𝑠 𝑂𝑝𝑡𝑖𝑚𝑎𝑙 and 𝜏𝑠
𝑂𝑝𝑡𝑖𝑚𝑎𝑙 are the optimal stay probability and sensing slot length,
respectively, while X is CM2M gateways total average energy cost for one data packet
transmission, including spectrum handoff, spectrum sensing, and data transmission. It is worth
noting that by applying energy detection, the energy consumption caused by the sensing
process can be calculated by the length of the sensing slot. By assigning a target false alarm
probability and a target detection probability (e. g. , ��𝑓 = 0.1 𝑎𝑛𝑑 ��𝑓 = 0.9) for each
channel, we can consider the time required to satisfy the sensing accuracy requirements while,
reducing this energy cost [13] as given below:
𝜏𝑠
=1
𝑦𝑓𝑠 𝑄−1(��𝑓) − 𝑄−1(��𝑑)√2γ + 1 (5.2)
5.3 Selective Antenna Sensing System Problem
By taking into consideration the challenges and the design features in the sections above, it can
be now present the proposed energy efficient mechanism to manage the available spectrum,
given in (5.1), and as follows:
𝑋(𝜏𝑠, 𝐿𝑠) = 𝑆𝐸𝑡 + N 𝜏𝑠𝐸𝑠 + 𝑁𝑃𝑥 𝐽𝑠𝑤 (5.3)
where 𝐸𝑡 and 𝐸𝑠 are the energy cost per second in Watts due to data transmission and sensing
respectively, and 𝐽𝑠𝑤 is the energy cost for one channel switching, in joules [13, 12].
Presuming that 𝐸𝑡, 𝐸𝑠 and 𝐽𝑠𝑤 are known for a given CM2M e-healthcare system, Ɲ is the
number of frames and the number of transmission slots of duration T seconds required to
complete one packet of transmission in the presence of a PU. 𝑆 is the time duration (in the unit
of seconds) of one data packet transmission without the PU’s presence, and 𝑃𝑥 is the probability
that the CM2M gateways decide to switch to a sensed-as-idle channel when the current channel
is sensed as busy and a minimum of, one of the other surrounding channels is sensed as free.
As shown in (5.1), the CM2M gateways minimum energy cost can be obtained by jointly
optimizing 𝜏𝑠 and 𝐿𝑠 under reliability thresholds. Despite this, it is not a convex optimisation
Page 5 - 6
problem under all thresholds, for a given 𝜏𝑠, the lower energy consumption can always be found
by employing the optimal 𝐿𝑠 as 𝐿𝑠 𝑂𝑝𝑡𝑖𝑚𝑎𝑙 = 𝐿𝑠
𝑢𝑝𝑝𝑒𝑟.
The reason is that function X becomes smaller as a function of 𝐿𝑠. Thus, the minimum value
of X happens at the farthest possible value of 𝐿𝑠. Because 𝐿𝑠 𝑂𝑝𝑡𝑖𝑚𝑎𝑙 = 𝐿𝑠
𝑢𝑝𝑝𝑒𝑟 , this reveals
that the probability of collision threshold is inactive in deriving 𝐿𝑠𝑢𝑝𝑝𝑒𝑟 [14]. Furthermore, the
probability of collision threshold is vital to get the shortest sensing slot, where 𝐿𝑠 minimum
(𝜏𝑠) is calculated from the probability of collision threshold. As a conclusion, the shortest 𝜏𝑠
required to meet all reliability thresholds is given by:
𝜏𝑠
= 𝑎𝑟𝑔𝜏𝑠min { 𝐿𝑠
𝑙𝑜𝑤𝑒𝑟(𝜏𝑠) ≤ 𝐿𝑠𝑢𝑝𝑝𝑒𝑟(𝜏𝑠)} (5.4)
where 𝐿𝑠 minimum (𝜏𝑠) is calculated by the collision threshold. Thus, the minimum 𝜏𝑠 needed
to meet collision threshold, detection probability and false alarm probability is given by:
𝜏𝑠
= max { 𝜏𝑠
, 𝜏𝑠} (5.5)
Note that 𝜏𝑠
, is calculated by the reliability thresholds: the probability of detection and the
false alarm probability from (5.2), while 𝜏𝑠
from the other reliability thresholds defined as
the target throughput, probability of collisions, and delay.
5.4 Problem Solution
Considering all the thresholds above, it will require longer 𝜏𝑠, to satisfy the conditions of
(𝐿𝑠𝑙𝑜𝑤𝑒𝑟 ≤ 𝐿𝑠
𝑢𝑝𝑝𝑒𝑟) ensuring a feasibility region of
𝐿𝑠, ∆ (𝜏𝑠) = (𝐿𝑠𝑢𝑝𝑝𝑒𝑟(𝜏𝑠) ≤ 𝐿𝑠
𝑙𝑜𝑤𝑒𝑟(𝜏𝑠) ≥ 0)
Page 5 - 7
and perhaps for practical values of 𝜏𝑠 no solution can optimise the existing problem.
Furthermore, it is justified by the fact that, for a given sensing slot duration and false alarm
probability, the probability of collision diminishes monotonically with any rise in the
probability of detection value. Therefore, any increase in detection probability leads to a
lower 𝐿𝑠𝑙𝑜𝑤𝑒𝑟 . Moreover, by enhancing the probability of detection, the feasibility region
∆ (𝜏𝑠) becomes obtainable with a shorter sensing slot time, which results in a decrease in
energy consumption caused by the sensing process. There are many schemes to improve the
probability of detection other than the typical Single Antenna Sensing (SAS), like multi-
antenna parallel sensing and cooperative sensing, but as the system model work under CM2M
concepts and goals it must avoid extra signalling to decrease cost and hardware complexity
when running a number of RF chains.
An ideal possible solution here comes by applying an Antenna Selection Sensing (ASS) scheme
to relieve the bounds of 𝐿𝑠. The ASS uses J antennas and Z RF chains (Z < J ) for the sensing
operation, where the CM2M gateways senses the channels for duration of time (𝜏𝑠).
Furthermore, the sensing slot of 𝜏𝑠 period is divided into J/Z sub-slots of period 𝜏𝑎𝑠 (e.g., 𝜏𝑠
= J/Z𝜏𝑎𝑠).
Z antennas are used to proceed with spectrum sensing in each sub-slot at the same time. For
instance, the system model can consider Z = 1 and J = 2. Compared to the multi-antenna
(conventional) spectrum sensing (e.g., J antennas are employed for sensing at the same time
as being employed for Z = K RF chains), the developed mechanism can decrease hardware
complexity and cost, since only one or a subset of RF chains are employed with a huge number
of antennas [14, 15].
5.5 System Simulation
The efficiency of the developed mechanism demonstrated using Matlab simulations. The
simulation value setting is given in Table 5.1; the numerical results are given here to prove the
analysis in the previous sections. The primary user is assumed to be a QPSK modulated signal
with a bandwidth of 7MHz. The sampling frequency is the same as the bandwidth of the
primary user. In the system model, the energy needed for spectrum sensing and sending data
assumed to be 𝐸𝑡 = 65 mW and 𝐸𝑠 = 45mW respectively.
Page 5 - 8
Table 5.1: Value Setting of Simulations
Parameter Value Description
ᶆ 6 Available channel number
S 6s Duration of one packet of data
T 0.8s Duration of each transmission
ρ 0.4 The probability of a channel being busy
��𝑑 0.9 The target probability of detection
��𝑓 0.1 The target probability of false alarm
γ -10 dB SNR
Jsw 40 (mJ) The energy consumption for one channel
switching
Figure 5.3 shows ASS (J = 2) can increase the detection probability which ultimately leads to
a lower probability of collision than SAS (J = 1) with the same value of 𝜏𝑠.
This can be explained by the relationship between the probability of collision, and the
probability of detection as the probability of collision value decreases with any increase in the
probability of detection and shows that ASS needs lower 𝜏𝑠 and energy consumption than the
traditional antenna (SAS) sensing when achieving similar probability of collision thresholds.
Moreover, for a given 𝐿𝑠, the probability of collision also monotonically gets smaller with 𝜏𝑠.
Again, this can be explained because a higher value of 𝜏𝑠 leads to a higher probability of
detection value and from the above we know, that will ultimately lead to decreasing the
probability of collision. Bear in mind the probability of collision is mainly caused by CM2M
gateways switching to channels with a primary user, due to failed detection.
Page 5 - 9
Figure 5.3: Using Antenna Selection Sensing (J = 2)
Figure 5.4 plots the 𝐿𝑠 bounds of the CM2M gateways as a function of sensing time 𝜏𝑠 for SAS
(J = 1), and ASS (J = 2). Using the same 𝜏𝑠, 𝐿𝑠𝑙𝑜𝑤𝑒𝑟 increased when an additional stringent
probability of collision thresholds Q (e.g., from Qz = 0.4 to Qz = 0.3 under J = 2) was used as
shown in Figure 5.4.
Such an increase in 𝐿𝑠𝑙𝑜𝑤𝑒𝑟, because of the additional stringent Q, allows the CM2M gateways
to remain on the current channel for a longer time to maintain the probability of collision
thresholds. The upper bound 𝐿𝑠𝑢𝑝𝑝𝑒𝑟 is decided by measuring which threshold (delay or
throughput) is more significant at any given time . Figure 5.4 considers the
throughput ( 𝜏𝑠, 𝐿𝑠𝑢𝑝𝑝𝑒𝑟) threshold. Bear in mind that any delay will yield a decrease in
average throughput, and therefore the CM2M gateway will need to switch channels repeatedly
to meet its required throughput thresholds The feasible region becomes available at 𝜏𝑠 = 16
(ms) if ASS (J = 2) is used. While using SAS (J = 1), the feasibility value can only achieve
at 𝜏𝑠 = 41 (ms).
This confirms, as stated in the previous section, that by using the ASS, the bounds of 𝐿𝑠 can be
relieved. Similarly, the energy consumption in CM2M gateways due to sensing can be
considerably reduced, as shown in Figure 5.5
Page 5 - 10
Figure 5. 4: 𝐿𝑠 Bounds of CM2M Gateways as a Function of 𝜏𝑠
Figure 5.5: The Energy Consumption of The CM2M Gateways Under Various Sensing
Mechanisms
Figure 5.5 shows the energy consumption of the CM2M gateways spectrum usage using ASS
and SAS sensing mechanisms. It demonstrates that the best stay probability for 𝜏𝑠 can be found
Page 5 - 11
at 𝐿𝑠𝑢𝑝𝑝𝑒𝑟. As explained before, the optimal sensing slot 𝜏𝑠
𝑂𝑝𝑡𝑖𝑚𝑎𝑙 is also calculated by the
reliability thresholds. As seen in Figure 5.5 the ASS offers reduced sensing time (𝜏𝑠= 16 (ms))
to meet all the thresholds, compared to SAS (𝜏𝑠 = 41 (ms)). Greater 𝐿𝑠 = 0.48 can be satisfied
while meeting all the thresholds with smaller 𝜏𝑠 by using ASS. As a total outcome, the energy
efficiency of the CM2M gateways can be improved significantly by using ASS under the same
thresholds.
5.6 Summary
In this Chapter, an energy efficient spectrum management mechanism for a CM2M e-
healthcare system has been designed. The designed mechanism simultaneously considers the
wait/switch trade-off in terms of channel switching probability, sensing /throughput trade-off
in terms of the period of sensing time and collision probability. Furthermore, the proposed
mechanism considers an ASS scheme for improved sensing accuracy and reduced energy
consumption. Simulation results show the optimality and efficiency of the used mechanism
with guarantees of the desired thresholds.
Page 5 - 12
References for Chapter 5
[1] Palicot and C. Roland.( 2005, October 5). “On the Use of Cognitive Radio for
Decreasing Electromagnetic-Radiation.” [online]: Available: https://hal.archives-
ouvertes.fr/hal-00776220.
[2] A. J. Jara, M. A. Zamora, and A. F. G. Skarmeta, “An architecture based on the
internet of things to support mobility and security in medical environments,” in
Proc. 7th IEEE Consumer Commun. Netw. Conf., Las Vegas, NV, 2010, pp. 1-5.
[3] W. Y. Chung, Y. D. Lee, and S. J. Jung, “A wireless sensor network compatible
wearable u-healthcare monitoring system using integrated ECG, accelerometer and
SpO2,” in Proc. 30th Annu. Int. Conf. Eng. Med. Biol. Soc., Vancouver, BC,
Canada, 2008, pp. 1529–1532.
[4] H. Furahata, "Electromagnetic interferences of electric medical equipment from
hand-held radiocommunication equipment," 1999 International Symposium on
Electromagnetic Compatibility (IEEE Cat. No.99EX147), Tokyo, 1999, pp. 468-
471.
[5] Stephen Wang, Filippo Tosato, and Justin P. coon, Toshiba research Europe limited,
“Reliable Energy-Efficient Spectrum Management and Optimization In Cognitive
Radio Networks: How Often Should We Switch”. IEEE Commun. Mag, Pages: 14
20.2014.
[6] A. Riker, T. Cruz, B. Marques, M. Curado, P. Simões and E. Monteiro, "Efficient
and secure M2M communications for smart metering," Proceedings of the 2014
IEEE Emerging Technology and Factory Automation (ETFA), Barcelona, 2014, pp.
1-7. doi: 10.1109/ETFA.2014.7005176.
[7] Michele Guerrini, Luca Rugini, and Paolo Banelli. The University of Perugia,
“Sensing-throughput tradeoff for cognitive radio networks,.”IEEE SPAWC,
Darmstadt, Germany, June 2013.
[8] C. F. Fung, W. Yu, and T. J. Lim, “Precoding for the multi-antenna downlink: Multi-
user gap approximation and optimal user ordering,” IEEE Trans. Commun., vol. 55,
no. 1, pp. 188–197, Jan. 2007.
[9] J. Lee and N. Jindal, “Symmetric capacity of MIMO downlink channels,” in Proc.
IEEE ISIT, Washington, DC, Jul. 2006, pp. 1031–1035.
Page 5 - 13
[10] D. N. C. Tse and P. Viswanath, Fundamentals of Wireless Communication.
Cambridge, U.K.: Cambridge Univ. Press, 2005.
[11] T. Yoo and A. Goldsmith, “Capacity of fading MIMO channels with channel
estimation error,” in Proc. IEEE Int. Conf. Commun., Paris, France, Jun. 2004, pp.
808–813.
[12] S.Alabadi, Predrag Rapajic, K. Arshad and Soheil Rostami, “Energy Efficient
Cognitive M2M Communications’’ International Journal of Interdisciplinary
Telecommunications and Networking (IJITN), Vol.8, Issue 3, July 2016.
[13] Wang, S., Wang, Y., Coon, J. P., & Doufexi, A. (2012). Energy-efficient spectrum
sensing and access for cognitive radio networks (pp. 13–17). Vehicular.
[14] S. Wang et al., “Antenna Selection Based Spectrum Sensing for Cognitive Radio
Networks,” IEEE PIMRC, Sept. 2011, pp. 364–68.
[15] Xiong, C.; Li, G.; Zhang, S.; Chen, Y.; Xu, S. Energy- and Spectral-Efficiency
Tradeoff in Downlink OFDMA Networks. IEEE Trans. Wirel. Commun. 2011, 10,
3874–3886.
Page 6 - 1
CHAPTER 6
ENERGY EFFICIENT SCHEDULING
ALGORITHM
Page 6 - 2
6.1 Introduction
Learning and selecting algorithms had a huge impact on cognitive radio and machine to
machine technologies. As mentioned in Chapter 2, a number of studies considered learning
algorithms for different CR networks. Panahi and Ohtsuki [1] presented a Fuzzy Q Learning
(FQL) based scheme for channel sensing in CR networks. Zhang et al. [2] presented a
reinforcement learning-based double action algorithm intended to increase the efficiency of
dynamic spectrum access in CR networks, in [3] channel allocation, Gallego et al. [4] presented
a game theoretic optimal solution for joint channel allocation and energy control in CR. A
number of studies employ partially observable Markov decision processes [5] and
reinforcement learning [6].
The main disadvantage of these techniques is their dependence on highly accurate reward
functions. Meanwhile, little effort has been made with regards to designing and developing a
CR learning engine with Experience Weighted Attraction (EWA) algorithms. EWA algorithms
[7, 8] give CR the ability to be aware of available channels characteristics online. By collecting
the history of channel statuses, it can foresee, select, and amend the best available channel,
dynamically test the quality of communication links, and eventually decrease system
communication outage probability.
The efficiency of this algorithm has been tested by the straightforward probability approach
[7] and with an EWA Handoff (EWAH) algorithm [105] in our preliminary studies. However,
again not much work has been done using EWA algorithms for more energy efficiency in
CM2M communications. This Chapter, utilise EWA algorithms to develop an EECS algorithm
for more energy efficient communications in the M2M healthcare system.
The rest of this Chapter is structured as follows: Section (6.2) demonstrates the system model
and explains our proposed algorithm. Simulation results and discussions are illustrated in
section (6.3), and Section (6.4) summarises the Chapter.
Page 6 - 3
Figure 6.1: Energy Efficient Scheduling Algorithm Flowchart
Page 6 - 4
6.2 Energy Efficient Scheduling Algorithm (System Model)
The system model considers a CM2M network is working with a number of CM2M gateways
and the primary link (Figure 6.1). As shown in Figure 6.1 the CM2M devices at the beginning
of transmission will choose to send data using elite channels, if the elite channels busy the
CM2M will choose to sleep or send data through another ideal channel. The process depends
on the throughput and delays thresholds if more data need to be transmitted CM2M devices
will choose elite or ideal channel to transmit if not, CM2M devices will go to sleep mode case
and save energy to prolong battery life.
The following assumptions considered up to the system requirements: the radio environment
is divided into ᶆ channels; detection is employed to sense the channels simultaneously, while
the CM2M gateways apply wideband sensing to test the availability status of all channels. In
the system, the secondary transmission assumed to be slotted in via periodic sensing at a
specific time, each frame consisting of a sensing slot of duration 𝜏𝑠 and a transmission slot of
duration T, where one of the ᶆ channels is assigned to one CM2M gateway.
𝑘𝑠𝑙 is the spectrum handoff delay, which will be added to the sensing slot in the CM2M
gateways frames, if the CM2M gateways decide to perform channel switching after addressing
the sensing outcome and other thresholds. The CM2M gateways at the start of each
transmission slot may then decide to send data on the current channel, switch to another free
ideal channel with a switch delay of 𝑘𝑠𝑙, or remain on the current channel without sending any
data; the primary user follows an on off traffic paradigm and is presumed to be continuous [9].
The average received SNR of the primary transition signal for all channels assumed to be the
same, at the same time the false alarm probability 𝑃𝑓 and the detection probability 𝑃𝑑
assumed to be one for each packet of data transmission [10]. The systems model assumes that
the received signal is sampled at sampling frequency 𝑓𝑠, and the frequency band with carrier
frequency 𝑓𝑐.
The goal is to reduce the energy cost for the CM2M gateways to send a packet of data under
both probabilities of false alarm and detection and constraints, as discussed in [11], proposing
that it can simultaneously optimize the sensing slot length 𝜏𝑠 and the stay probability on the
Page 6 - 5
ideal current channel 𝐿𝑠 so that the power cost of one complete data packet transmission is
reduced as follows:
( 𝜏𝑠𝑂𝑝𝑡𝑖𝑚𝑎𝑙, 𝐿𝑠
𝑂𝑝𝑡𝑖𝑚𝑎𝑙) = min 𝜏𝑠 ,𝐿𝑠
𝑋 (𝜏𝑠, 𝐿𝑠) (6.1)
where 𝐿𝑠 𝑂𝑝𝑡𝑖𝑚𝑎𝑙 and 𝜏𝑠
𝑂𝑝𝑡𝑖𝑚𝑎𝑙 are the optimal stay probability and sensing slot length,
respectively, while X is CM2M gateways overall average energy consumption for one data
packet transmission, which includes spectrum sensing, spectrum handoff, and data
transmission.
However, by employing staying probability, the energy cost due to sensing is calculated by the
duration of the sensing slot. By allocating a target false alarm probability and a target detection
probability (𝑒. 𝑔. , ��𝑑 = 0.1 𝑎𝑛𝑑 ��𝑑 = 0.9) for each channel, it can consider the duration
required to satisfy the sensing threshold requirements while reducing the energy consumption
[11] as given below:
𝜏𝑠min =
1
𝑦𝑓𝑠 (𝑄−1(��𝑓) − 𝑄−1(��𝑑)√2γ + 1 (6.2)
By studying the design characteristics and challenges in the equations above, it can now
develop an energy efficient mechanism (Figure 6.1) to select the best available channels, given
in (1) and as follows:
𝑋(𝜏𝑠, 𝐿𝑠) = 𝑆𝐸𝑡 + Ɲ 𝜏𝑠𝐸𝑠 + Ɲ𝑃𝑥 𝐽𝑠𝑤 (6.3)
where 𝐸𝑡 and 𝐸𝑠 are energy consumption per second in Watts due to data transmission and
sensing, and 𝐽𝑠𝑤 is the energy consumption for one channel switching, in joules [12].
Page 6 - 6
Presuming, that 𝐸𝑡, 𝐸𝑠 and 𝐽𝑠𝑤 are known for a given CM2M healthcare system, Ɲ is the
number of frames and the number of transmission slots of time T seconds needed to finish one
packet of transmission in the existence of primary transition, while 𝑆 is the time duration (in
the unit of seconds) of a single data packet of transmission without the primary transmissions
existence, and 𝑃𝑥 is the probability that the CM2M gateways will choose to switch to an ideal
channel when the current channel is busy and at least one, of the other channels is free and its
function of 𝐿𝑠 and 𝜏𝑠 [13, 14].
As shown in (6.1), the CM2M gateways low energy consumption can be gained by
simultaneously optimizing 𝜏𝑠 and 𝐿𝑠 under reliability thresholds. Despite this it is not a convex
optimisation problem under all constraints, for a given 𝜏𝑠, the minimum energy cost can be
found by utilising the optimal 𝐿𝑠 as 𝐿𝑠 𝑂𝑝𝑡𝑖𝑚𝑎𝑙 = 𝐿𝑠
𝑢𝑝𝑝𝑒𝑟.
On the other hand, in such a CM2M system the idle probability of channel К (1 ≤ К ≤ 𝑛) can
be expressed as 𝛼𝑘, or 𝛢𝐴 = [𝛼1, 𝛼2, . . . , 𝛼𝑛−1, 𝛼𝑛] in vector form. Let 𝛽К be the successful
transmission probability of channel К (1 ≤ К ≤ 𝑛); then 𝐵𝐵 =
{𝛽1, 𝛽2, . . . , 𝛽𝑛−1, 𝛽𝑛}. Perhaps the radio channel features vary by time; the channel idle
probability and effective transmission probability of channel К (1 ≤ К ≤ 𝑛) should not be
one at various time 𝑡; then the forms of probabilities after inserting time parameter 𝑡 are
𝛢𝐴(𝑡) = {𝛼1(𝑡), 𝛼2(𝑡), . . . , 𝛼𝑛−1(𝑡), 𝛼𝑛(𝑡)}
and
𝛣𝐵(𝑡) = { 𝛽1(𝑡), 𝛽2(𝑡), … , 𝛽𝑛−1(𝑡), 𝛽𝑛(𝑡) }
To decrease the level of complexity of the channel selecting strategy in the CM2M healthcare
system, exponential functions should first be avoided. Next, the system model considers an
algorithm which should facilitate the calculation procedure and optimize and update the
objective operation to provide better energy efficiency and better throughput. Furthermore, the
system model defines the probability of channel selecting 𝑦 in the channel preferable selection
policy 𝑆К𝑦
at time 𝑡 as 𝑃К𝑦
(𝑡) then the mathematical expression of 𝑃К𝑦
(𝑡) is
Page 6 - 7
𝑃К𝑦(𝑡 + 1) =
1- σ1-σ. {1-I[𝑆К
𝑦,𝑆К(𝑡)]}
(6.4)
Where
x( К ) = {1, 𝑇𝑟𝑎𝑛𝑚𝑖𝑠𝑠𝑖𝑜𝑛 𝑓𝑎𝑖𝑙𝑢𝑟𝑒 𝑜𝑛 𝑐ℎ𝑎𝑛𝑛𝑒𝑙 𝑦, 0, 𝑆𝑢𝑐𝑐𝑒𝑠𝑠𝑓𝑢𝑙 𝑡𝑟𝑎𝑛𝑚𝑖𝑠𝑖𝑜𝑛 𝑜𝑛 𝑐ℎ𝑎𝑛𝑛𝑒𝑙 𝑦,
𝜋К[𝑆К𝑦
, 𝑆−К(𝑡)] = {0, 𝑐ℎ𝑎𝑛𝑛𝑒𝑙 𝑦 𝑖𝑠 𝑠𝑒𝑛𝑠𝑒𝑑 𝑏𝑢𝑠𝑦 1, 𝑐ℎ𝑎𝑛𝑛𝑒𝑙 𝑦 𝑖𝑠 𝑠𝑒𝑛𝑠𝑒𝑑 𝐼𝑑𝑙𝑒
(6.5)
and 𝐼[⋅] is the indicator function, which is defined as follows:
I(x,s) = {1, 𝑥 = 𝑠,0, 𝑥 ≠ 𝑠,
(6.6)
Parameters 𝜎 and µ are attenuation coefficients of probability and 𝜎 < µ ∈ (0, 1). From the
analysis of (6.4) it’s clear that in the CM2M environment sensing time, when knowing the
current status of channel 𝑦 as being occupied (primary user making transition), the status
becomes 0 (unavailable), and the scheme of choosing channel 𝑦 as a transmission channel
will get no gain, or the award operation value of 𝜋К[𝑆К𝑦
, 𝑆−К(𝑡)] is 0 and the channel selecting
probability declines to (1 − µ) 𝑃К𝑦(𝑡); while knowing the current status of channel 𝑦 as being
idle (no PU user transmission in the channel ), the status changes to 1 (available), and the
scheme of choosing channel 𝑦 as a transmission channel will get the gain of 𝜋К[𝑆К𝑦
, 𝑆−К(𝑡)]
Page 6 - 8
respectively. Furthermore, the value of 𝜋𝑦[𝑆К𝑦
, 𝑆−К(𝑡)] is presumed to equal 1 and the channel
selection probability is updated to (1- µ). 𝑃К𝑦(𝑡) + µ.
The free available channels are elite channels for CM2M gateways transmission, and the elite
channel with the biggest 𝑦 probability of channel selecting will be chosen for sending data. If
more than one channel reaches the biggest selection probability, then one of these channels will
be selected at random after successful transmission, the channel selecting probability will go
up to (1 − σ).[( 1 − µ). 𝑃К𝑦(𝑡) + µ] + σ. But, if the transmission fails, the channel selecting
probability will update to (1 − σ).[( 1 − µ). 𝑃К𝑦(𝑡) + µ].
6.3 Simulation and Results
This section demonstrates the efficiency of our algorithm EECS using Matlab simulation. The
simulation value settings are listed in Table 6.1; the outcomes are given here to confirm the
analysis in previous sections. The primary transmission is presumed to be a QPSK modulated
signal with a bandwidth of 8MHz while the energy needs for spectrum sensing and
transmission are 𝐸𝑡 = 69.5 mW and 𝐸𝑠 = 40mW respectively.
Since the value of 𝜎 should be lower than parameter µ, the coefficients µ assumed to be at
initial value 0.1 due to comprehensive experience. While there must be a number of variations
between each channel, the idle probabilities of such channels will not be one. To reflect the
typical channels’ available probabilities, a symmetrical distribution vector in the range of 1 to
0 will be picked for the idle probability of each channel; that is, the first channel idle probability
vector AA0 = {0.4, 0.9, 0.6, 0.5, 0.7}, while the first channel successful transmission
probability vector BB0 = {3/4, 8/9, 5/6, 4/5, 6/7}.
Then the first channel available probability vector Γ0 = AΑ0 ⋅ BΒ0 = {0.3, 0.8, 0.5, 0.4, 0.6}
In order to check that this smart algorithm is capable of deciding and guiding CM2M gateways
real-time switch to the new transmission channel with the highest selecting probability online
precisely, the channel idle probability vector will update to AΑ1 = {0.6, 0.4, 0.7, 0.9, 0.5} and
the channel successful transmission probability vector will change to BΒ1 = {5/6, 3/4, 6/7,
8/9, 4/5} after 40 rounds through the simulation.
Thus, the channel available probability vector will be Γ1 = AΑ1 ⋅BΒ1 = {0.5, 0.3, 0.6, 0.8, 0.4}
after the simulation circumstances change.
Page 6 - 9
Table 6.1: Value Settings of Simulations
Parameter Value Description
ᶆ 5 Available channel number
S 6s Duration of one packet of data
T 0.8s Duration of each transmission
γ -10dB SNR
𝐽𝑠𝑤 40mJ The energy consumption for one channel
switching
ρ 0.4 The probability of a channel being busy
��𝑑 0.9 The target probability of detection
��𝑓 0.1 The target probability of false alarm
CM2M 2 Cognitive Machine-to-Machine gateways
Taking randomness of the parameters above through a practical wireless network into account,
the numbers formed in each simulation round meet exponential distribution of the
corresponding parameter above accompanied by the general rule.
Next, the system model applies a simple repeated experimental method to check the
effectiveness of the probability of our EECS algorithm. That is, Turn Based Strategy (TBS), a
single uniformly distributed random number within the range (0, 1), is generated in each round.
If this number is less than the channel available probability 𝛼К, channel К is judged as being in
an idle available state; otherwise it is in a busy unavailable state.
The idle channel with the highest selection probability will be the preferred communication
channel in the current round. If more than one channel reaches the highest probability of
channel selecting, then one of these channels will be selected randomly. After the algorithm
selects the preferred channel 𝑦, a single uniformly distributed random number within range (0,
1) is also generated. Communication channel transmission is successful if this number is less
than the probability of successful data transfer completion 𝛽𝑦; otherwise it, fails.
Page 6 - 10
After the parameters above are set, the outcomes of the channel selecting probability depending
on the EECS algorithm are demonstrated in Figure 6.2. Using channel selecting probabilities,
following a short establishing process, the EECS learning algorithm will typically choose and
track channel 1 as the initial gateway channel, but its selection probability fluctuates slightly
around 0.73.
Due to optimal channel availability, the probability changes after the 43rd round and the
selection probability of channel 1 drops tremendously, while the selection probability of
channel 2 grows, slowly but surely overtaking the selection probability of channel 1 after 48
rounds. Channel 2 finally takes over from channel 1 to become the optimal gateway channel
under the new channel available probability status, and channel 1 will be the second in the list.
Figure 6.2: The Outcomes of Channel Selecting Probability Based on EECS Algorithm
Figure 6. 3 shows the staying probability increase ( 𝐿𝑠 ) when EECS algorithm used, the reason
being that the quality of the selected channels will be better which will lead to less switching
of CM2M gateways in the available channels as only the preferred channels with fewer PU
transmissions will be selected.
Page 6 - 11
Figure 6.4, shows the energy cost of the CM2M gateways spectrum usage under the EECS
algorithm. It proves that the optimal energy solution can always be found at the optimal
maximum value of 𝐿𝑠. Because the higher probability of staying means less switching of
CM2M gateways in the available channels. At the same time, better throughput could be
achieved leading to less 𝜏𝑠 (less energy consumption through sensing process) and better
energy efficiency in the system overall.
Figure 6.3: Staying probability at 𝜏𝑠 Optimal
Page 6 - 12
Fig 6.4: Energy Consumption at 𝜏𝑠 Optimal
6.4 Summary
This Chapter proposes an energy-efficient channel is selecting algorithm. The proposed
algorithm improves CM2M gateways channel selection and the probability of staying by at
least (12%) due to the increase in transmission channel quality which, leads to less switching
of CM2M gateways between the available channels. Simulation results demonstrate the
optimality and efficiency of the algorithm with the guarantees of the desired constraints.
Page 6 - 13
References for Chapter 6
[1] F. H. Panahi and T. Ohtsuki, “Optimal channel-sensing scheme for cognitive radio
systems based on fuzzy q-learning,” IEICE Transactions on Communications, vol.
97, no. 2, pp. 283–294, 2014.
[2] Y. L. Teng, F. R. Yu, K. Han, Y. F. Wei, and Y. Zhang, “Reinforcement-learning-
based double auction design for dynamic spectrum access in cognitive radio
networks,” Wireless Personal Communications, vol. 69, no. 2, pp. 771–791, 2013.
[3] B. F. Lo and I. F. Akyildiz, “Reinforcement learning for cooperative sensing gain
in cognitive radio ad hoc networks,” Wireless Networks, vol. 19, no. 6, pp. 1237-
1250, 2013.
[4] Gallego, J. R., M. Canales, and J. Ortin, Distributed resource allocation in cognitive
radio networks with a game learning approach to improve aggregate system
capacity, Ad Hoc Networks, Vol. 10, Issue 6, 2012, pp. 1076-1089.
[5] Torkestani, J. A. and M. R. Meybodi, A Learning Automata-Based Cognitive Radio
for Clustered Wireless Ad-Hoc Networks, Journal of Network and Systems
Management, Vol. 19, Issue 2, 2011, pp. 278-297.
[6] Shan-Shan, W., et al., Primary User Emulation Attacks Analysis for Cognitive
Radio Networks Communication, TELKOMNIKA Indonesian Journal of Electrical
Engineering, Vol. 11, Issue 7, 2013, pp. 3905-3914.
[7] Y. Sun and J.-S. Qian, “Cognitive radio channel selection strategy based on
experience-weighted attraction learning,” TELKOMNIKA Indonesian Journal of
Electrical Engineering, vol. 12, no. 1, pp. 149–156, 2014.
[8] Y. Sun and J. S. Qian, “EWA selection strategy with channel handoff scheme in
cognitive radio,” Sensors & Transducers, vol. 6, pp. 68–74, 2014.
[9] F. H. Panahi and T. Ohtsuki, “Optimal channel-sensing scheme for cognitive radio
systems based on fuzzy q-learning,” IEICE Transactions on Communications, vol.
97, no. 2, pp. 283–294, 2014.
[10] Research and Markets: Machine-to-Machine (M2M) Communication in Healthcare
2010-20: Reviews the Major Drivers and Barriers to Growth of M2M.
http://dx.doi.org/10.1787/5k9gsh2gp043.
Page 6 - 14
[11] S.Alabadi, Predrag Rapajic, K. Arshad and Soheil Rostami, “Energy Efficient
Cognitive M2M Communications’’ International Journal of Interdisciplinary
Telecommunications and Networking (IJITN), Vol.8, Issue 3, July 2016.
[12] Liang, Y. C., Zeng, Y., Peh, E. C., & Hoang, A. T. (2008). Sensing-throughput
trade-off for cognitive radio networks. IEEE Transactions on Wireless
Communications, 7(4), 1326–1337. doi:10.1109/TWC.2008.060869.
[13] Wang, S., Wang, Y., Coon, J. P., & Doufexi, A. (2012). Energy-efficient spectrum
sensing and access for cognitive radio networks (pp. 13–17). Vehicular.
[14] D. Zhang, K. Li, and. Xiao, “An improved cognitive radio spectrum sensing
algorithm,” TELKOMNIKA Indonesian Journal of Electrical Engineering, vol. 11,
no. 2, pp. 583–590, 2013.
Page 7 - 1
CHAPTER 7
THE FUTURE OF M2M IN THE
HEALTHCARE SECTOR
Page 7 - 2
7.1 Introduction
A dramatic shift in the business models of hospitals and healthcare systems is evident as we
enter the 21st century, and at the heart of this shift is Machine-to-Machine technology [1]. This
new technology marries the use of medical machines and wireless communication systems to
offer new possibilities and application areas for the monitoring of both symptoms and diseases
[1, 2]. The global telehealth market is soaring as the use of remote monitoring technology
moves into the mainstream (Figure 7.1).
Figure 7.1: Telehealth Usage [2]
One of the primary objectives of the industry is to keep people from occupying hospital beds
wherever possible, and occupancy rates are in fact slowly declining [3]. Inpatient hospital
admissions dropped from an average of 123.2 per thousand people in 1991 to 111.8 per
thousand in 2011, highlighting the switch from inpatient to outpatient care [4].
Medical bodies need to discharge patients as quickly as possible to maximize the number of
beds available. Great care must be taken when selecting patients for release, as the Affordable
Care Act stresses that medical bodies will be penalized if discharged patients are re-admitted
within a specific timeframe. The best approach to minimize the chance of re-admission is
Page 7 - 3
through the implementation of M2M healthcare applications such as remote health monitoring
[6] and smart sensors.
7.2 The Factors Driving M2M Adoption
The twin factors of an aging population (the post-war ‘Baby Boomers’) and an increase in
chronic illnesses combine to put pressure on the health industry and its ability to monitor
patients. The simple solution of remote monitoring for such patients will greatly assist in
relieving this pressure. Fortunately, the appetite for wireless technology by the general public
- and in particular the widespread dependence on mobile phones, and the use of these for
purposes other than just communication has accelerated public acceptance of M2M healthcare
applications. Among those surveyed by the Economist Intelligence Unit (EIU) in 2012, 48%
of individuals stated that they believed mobile healthcare applications could increase the
quality of services in the medical sector [6, 7, 8].
In addition, the economic benefits are considerable. In 2012, approximately 308,000 patients
worldwide were remotely monitored for conditions such Chronic Obstructive Pulmonary
Disease (COPD), Congestive Heart Failure (CHF), Hypertension, Diabetes, and mental health
conditions, according to a report from InMedica [128]. This number is predicted to increase to
1.8 million patients in 2017 [9]. The growth in remote monitoring is predicted to save the
medical sector worldwide up to $36 billion by 2018, according to a projection by Juniper
Research [10]. The advantages are not just economical. Mareca Hatler, Director of Research at
ON World, confirms the advancements of M2M healthcare applications ”In addition to
reducing costs, cloud-connected wireless sensing solutions are improving the quality of
healthcare services as well as supporting the latest innovations for aging in place, self-
management of chronic conditions, and general wellness" [11].
7.3 M2M Opportunities in the Healthcare Sector
Currently, there are eight major healthcare application groups [7, 8, 9, 12]:
1. Home monitoring: Involves patients self-testing, which means home monitoring can
become a treatment option. Data from the medical devices are then transmitted to healthcare
providers for disease management. Some of the most common conditions being supervised
Page 7 - 4
today are chronic diseases including cardiac arrhythmia, hypertension, ischemic diseases,
sleep apnea, diabetes, hyperlipidemia, asthma, and COPD [12]. These conditions are costly
to the healthcare service and reduce both life expectancy and quality of life. The use of
information and communication technologies can lead to reduced costs and better medical
delivery.
2. Clinical monitoring: Improves healthcare management by safely decreasing patients’
hospital stays and visits. Patient sensors also assist doctors, helping them to spot early
warnings of medical deterioration and enabling them to apply treatment earlier than
physical diagnosis would allow. These solutions dramatically improve quality of life by
helping patients regain their mobility and independence [13].
3. Telemedicine: Especially useful in rural areas, Telemedicine helps reduce the high costs
of serious illnesses by allowing doctors to oversee the condition of several patient cases
each day, eliminating the need for unnecessary visits. Portable, wearable, and even
implantable sensors and tools may be used as tracking systems to determine a patient’s
location. Monitoring systems may be implemented to constantly scan vital signs and
provide vital data to healthcare providers or, in case of an emergency, automatically send
an alert to a doctor or healthcare facility. A real-time information system could mean the
difference between life and death in case of heart failure, diabetic comas, and other serious
illnesses [10].
4. First responder connectivity: In an emergency situation, every second counts and
equipment have to be reliable, without exception. Unshakable reliability and innovation are
required to deliver this service at its fullest potential [14].
5. Connected medical environments: The burgeoning fitness industry has led end-users to
the point of making their own informed decisions on health and fitness programs. M2M
solutions can not only be used to oversee vital signs during exercise by analyzing data
obtained from connected heart-rate monitors and other devices but can also make real-time
transmissions of the data. This enables users to keep track of their health and fitness
progress and offers them the chance to share their workouts on social networks.
6. Clinical remote monitoring: Costly home visits to patients with chronic conditions may
be reduced to a minimum by the use of clinical remote monitoring, whereby devices may
be fitted to or used by the patient to enable continuous monitoring. For example, Silent
Page 7 - 5
Observer, developed by Sukrut Systems, uses technology to police providers’ activity on
ultrasound machines. The data is collected, reported daily to local agencies, and cross-
checked by those agencies to make sure the providers have filled out mandatory pregnancy
reporting forms. The use of this intelligence may aid in decreasing the incidences of illegal
female foeticide [16].
7. Assisted living and clinical trials: Patient tracking systems and monitoring devices may
be used to ensure the health and safety of patients in the absence of caregivers. It also serves
to assist the disabled and elderly with their daily tasks, helping them live independently in
their own homes [17].
8. Asset management: This is used to track and show the availability of mobile healthcare
equipment at any time, to schedule routine cleaning and maintenance tasks, and reduce costs
with equipment safety monitoring and incident tracking [18].
7.4 Real World Examples
In the UK alone, nearly 1 million people have been diagnosed with heart failure, and this figure
is growing by almost 60,000 new cases each year [4]. Continuous monitoring used to be
unachievable, and health alerts (such as irregular heartbeats) were flagged up only after data
from a body-worn external controller unit could be transmitted to clinicians via a local internet
connection, causing delays and limiting patients’ mobility [11, 12].
Figure 7.2: HeartAssist5, Numerex [19]
Page 7 - 6
Among other businesses who are working to introduce improvements, a company called
Numerex has developed “The Reliant Heart Assist Remote Monitoring System, HeartAssist5”
(Figure 7.2), an M2M-enabled device which allows patients to monitor their heart health not
only when at home, but also while traveling [19]. The system utilizes the secure, cloud-based,
fully integrated fast platform. At the hub of the system is the HeartAssist5 Conquest Controller.
At present, the device is being used by more than 110,000 patients, continuously receiving
information from the monitoring system.
This information is then transmitted in packets to a controlled and safe data center monitored
by trained supervisors in the same way as for domestic security systems. Physicians have
remote access to this data whenever required, and in critical situations, the device sends an alert
to the patient's caregivers, giving them the necessary information to arrange meetings with the
patient’s physician or to make arrangements for hospital admission.
The HeartAssist5 is currently available in Europe, while in the US, it is undergoing evaluation
by the Food and Drug Administration (FDA). Expected benefits, apart from patient
reassurance, would be increased freedom and peace of mind for the caregiver, and better use
of the healthcare system's resources.
Benefits are not restricted to patients alone; the stress of caring for a sick or elderly family
member should not be under estimated and can seriously impact on the day-to-day life of
caregivers. Many family members take time away from their jobs to care for sick relatives
because they worry about what might happen when they are left at home by themselves.
However, heart care home applications will monitor patients and send all the required
information to the dedicated emergency services by exploiting wearable technology that tracks
their condition.
These wearables range from pulse detectors and blood pressure monitors worn around the wrist
to a personal emergency response framework worn as a necklace, to sleep apnea machines
worn on the face. These devices report health information and statistics through mobile
networks and other internet services to providers and health entities. This flow of
information/data signals actual and possible healthcare emergencies, as well as detected
abnormalities which should be dealt with immediately. The signals would travel via email or
mobile alerts to doctors and family members/caregivers or summon an emergency medical
technician, depending on the urgency of the situation [20].
Page 7 - 7
Moreover, M2M healthcare home applications boost the level of self-sufficiency for the elderly
and chronically ill by letting them proceed with their daily routines safe in the knowledge that
they will be able to summon help in the case of an emergency [21, 22]. In forthcoming years,
a significant number of patients’ houses will have intelligent machine networks to remotely
monitor their activities, allowing those living on their own to call for help when necessary.
Machines can be placed in cabinets, living rooms, and any other part of the home. Research
carried out by Orange has proved that 73% of senior citizens in EU would feel safer with a
‘tele-assistance’ tool which could recognize problems and call for assistance in case of an
emergency. M2M can be critical here, with devices like motion detectors in chairs and beds
able to send off alerts after periods of non-movement – or in the event of unwanted movement
already utilized in a number of markets nowadays [11].
Nowadays, the similar technology exists in the market thanks to innovators such
as SimplyHome who are pioneering home M2M practices. The systems from SimplyHome
monitor day-to-day activity and can alert caregivers or health professionals if no routine
activity has taken place, or if something unusual has occurred in a patient's home [23].
GTX Corp, another innovator in remote patient monitoring machines, offers services to
remotely supervise patients when they leave their homes. Their product, GPS SmartSole
(Figure 7.3) is a GPS tracking shoe sole for those with Alzheimer’s disease, dementia, and
other mental issues. A carer or family member can create a geofence (e.g., designated
boundaries placed around the patient's house or neighborhood) [ 24, 25].
Whenever SmartSole wearer crosses a geofence, family members and healthcare professionals
will get an alert explaining the situation, which will help them take quick action to physically
locate the patient, so the patient can either go back within the set geofence or return home or
call for emergency services if need be.
Page 7 - 8
Figure 7.3: SmartSole, GTX Corp device [25]
Another challenge for patients, especially for those with dementia, Alzheimer's disease, or with
a complex drug regime, is keeping track of their prescribed medicine routine. If a patient under
or overdoses, it can lead to serious implications and possible health emergencies. M2M enabled
“smart pill boxes,” and packaging makes it simple for health bodies and caregivers to keep
track of a patient’s medicine intake and schedule [26]. Already in use today, pillboxes fitted
with embedded sensors can tell if a pill has been removed from a bottle or packet and send a
report to let the health care provider know whether the patient has been taking the correct
dosage of their medication at the correct time [27].
This is instrumental in preventing accidental overdose. Furthermore, the pillboxes can be
programmed to send an alert to family members and caregivers if the medicine has not been
taken [28]. MedSignals’ have developed their M2M MedSignals Pill Case Device [29]. The
Pill Case has 4 compartments, each capable of holding up to 70 small pills. The compartments
remain closed until it is time for the patient to take the medication, at which point one of four
corresponding buttons will light up, so patients know which compartment to open. MedSignals
Pill Case also has voice devices (speakers) and a small monitor on board so that other directions
can be played to clarify which pills to take along with the correct dosage required.
Other companies are taking a different approach to medicine accountability solutions by
incorporating M2M technology directly into the medication’s original packaging. Mevia, for
example, has developed smart packages that automatically send alerts whenever a pill is
removed from a pack or bottle. Mevia’s packaging method spares the patient or carer the task
of having to remember to remove pills from their original packaging and put them in a separate
Page 7 - 9
smart container - especially useful for patients with poor memory or confusion, and for
caregivers with a busy schedule. In the next few years, pharmaceutical companies may start
partnering with companies like Mevia to help patients remember to take their medication or
notify caregivers to give medication and assure the right amount or dose is being taken every
time.
The December 2013 mHealth Summit - the largest assembly of its kind concerned solely with
the fusion of mobile and healthcare technologies saw the unveiling of several state-of-the-art
devices currently in development. Some examples of future innovations include [10, 30]:
• A pendant with patented automatic fall detection algorithms which can send an alert in
case of personal emergency;
• A bottle cap digitally enabled to allow for compliance monitoring and able to alert
patients when medication is due, which may be fitted to standard prescription bottles;
• A service which will enable a mobile phone to automatically request a "house call"
service, which brings medical assistance to the patient within two hours.
These are only a few examples. Other, further advanced, technologies are forecasted to be
deployed within the next few years. A study released by the research company ON World
estimates that 18.2 million health and wellness Wireless Sensor Networks (WSNs) are
expected to be shipped globally in 2018, and this will generate around $16.3 billion in annual
revenue. Another trend that is predicted to rise to prominence very quickly is the use of
disposable body-worn wireless medical sensors.
7.5 M2M Applications: Future Challenges
The public has seen just the start of the M2M evolution in healthcare. Allied Business
Intelligence (ABI) has predicted a huge development of more than 5 million disposable MBAN
sensors being shipped by 2018, and even this is a tiny proportion of its potential [31].
Meanwhile, Machina Research has predicted a global installation of 847 million M2M
connected health machines by 2023, totaling a price tag of $91 billion. North America will be
the largest consumer of M2M technology during this time, with 386 million M2M machines
by 2023. Similar numbers will follow in Asia and the EU. These projections cover the whole
Page 7 - 10
spectrum of remote health monitoring, connected medical environments, and improved assisted
living aids.
Nonetheless, healthcare technology needs to be reliable, as malfunctions are not just frustrating
but could be life-threatening. For M2M to succeed in healthcare applications, essential
requirements must include the following: good quality and authentic communications in both
very local wireless systems and national and international mobile networks; small, robust and
user-proof hardware; good battery life without replacement; and secure and reliable
communications.
Realistically, M2M in the healthcare sector could face extra challenges such as the following:
• Regulations that allow and promote development in sectors such as energy and
automotive industries but are a hindrance to innovation in the healthcare sector. For
instance, many healthcare practices are funded according to the number of patients
they see, which doesn’t encourage them to prioritize the use of M2M applications
that solve patient problems without the need for a hospital visit;
• Trying to identify business models that harmonize M2M technology with existing
healthcare incentive schemes;
• Lengthy lead times required to win regulatory approval for new devices which will
operate in a highly complex ecosystem, connecting doctors, hospitals, ambulances,
and care homes; and
• Concerns regarding customer privacy and data security.
7.6 Security Risk and Vulnerabilities in Healthcare Sector
Cybersecurity is always of paramount importance. In 2017, the UK National Health Service
was hit by a cyber-attack, causing major problems for a number of hospitals and affecting a
large number of patients [32, 33]. This demonstrates how important it is to understand how
security issues and vulnerabilities manifest themselves in M2M health devices. In this Chapter,
a number of the security challenges are explored and presented as the following:
Page 7 - 11
• Products are rushed on to the market – Security is usually the last concern of many
manufactures due to a tight timescale. Devices are often released by the manufacturers
having undergone little or no security assurance testing [28].
• Small form factors and limited capability components – Healthcare CM2M devices
are generally made to be very small, with the unfortunate consequence of being
restricted to using components with limited capability. For example, slots and memory
have limits on the level of security or encryption that they can provide. These limits are
usually not in line with current best practices and are exploitable through existing and
known vulnerabilities.
• Lack of, or vulnerable, secure update mechanism – In the event of security
vulnerabilities being detected in healthcare M2M devices, coping with the security risk
and assuring successful updates to all compromised machines may not be a simple
mission. It would not be feasible to entrust the end-user to periodically check for and
install updates in, for instance, an implanted pacemaker or a connected insulin pump;
and it is, in fact, questionable as to whose responsibility this actually would be - the
vendor, the manufacturer, or the hospital/medical centre overseeing the patient’s
treatment?
However, new and efficient security schemes need to be deployed in the near future to achieve
secure and reliable communications in M2M healthcare applications. A number of significant
requirements needed to improve security in the M2M communication of healthcare
applications are shown in table 7.1. Meeting these requirements will significantly boost
security in healthcare systems and at the same time will encourage hospitals and other medical
bodies to trust healthcare M2M applications and use them more widely [10, 18].
Page 7 - 12
Table 7.1: CM2M Communications Security Requirements in Medical Applications Sector
Requirement name Requirement explanation
Non-Repudiation Non-repudiation implies that a node cannot
deny sending a message or data sent earlier
Authorization & Authentication
Authorization & Authentication allows an
M2M healthcare machine to assure the
identity of the peer with which it is
communicating
Data/Key Freshness
Because M2M healthcare networks provide
some time-varying measurements, data or
key freshness ensures that each data set is
recent, and no adversary can replay old
messages
Resiliency
If medical care machines are compromised,
alternative security approaches should still
protect the data from any hacking
Self-healing
If a healthcare machine in an M2M healthcare
network fails or runs out of power, then
collaborating machines should provide
second-line security and maintain minimum
security levels
Confidentiality
Confidentiality ensures that unauthorized
users are denied accessibility to medical
information, while confidential messages are
protected from eavesdroppers
Page 7 - 13
7.7 Summary
Healthcare mobile phone devices and applications will eventually become the standard for both
caregiver and patient, while real-time technology (e.g., video) will become a significant part of
most medical care management. The massive development of healthcare systems will face
many challenges in terms of security, spectrum availability, and battery life limitations.
Cognitive radio and smart technologies will solve some of these challenges, as explained in the
previous chapters. However, new technologies and schemes need to be developed to cope with
all M2M communications challenges and to enable efficient, reliable, and secure
communications among all M2M applications in the medical sector.
Page 7 - 14
References for Chapter 7
[1] Wayne Caswell, " Global Telehealth Market to Expand 10x by 2018 " [Online:
Accessed16/06/2017]:http://www.mhealthtalk.com/2014/01/global-telehealth-
market-to-exp -and-10x-by-2018/.
[2] M. Malik, “Heart rate variability: Standards of measurement, physiological
interpretation, and clinical use,” Circulation, vol. 93, no. 5, pp. 1043–1065, Mar.
1996.
[3] Aditi Pai, " 51 digital health metrics in 2013 ", [Online: Accessed 16/06/2017]:
http://mobihealthnews.com/27638/51-digital-health-metrics-in-2013/ no link in the
text.
[4] Health and Social Care Information Centre, Hospital Episode Statistics, Admitted
Patient Care, England - 2014-15, HSCIC, November 2015.
[5] Emerging healthy path for growth, [Online: Accessed 20/06/2017]:
http://www.pwc.com/mx/es/industrias/archivo/2012-06-emerging-mhealth.pdf.
[6] M. Kuroda and M. Fukahori, "Affordable M2M enabled e-health using standard ban
technology," 2014 IEEE Healthcare Innovation Conference (HIC), Seattle, WA,
2014, pp. 276-279.doi: 10.1109/HIC.2014.7038928.
[7] P. S. Pandian, K. Mohanavelu, K. P. Safeer, T. M. Kotresh, D. T. Shakunthala,
P.Gopal, and V. C. Padaki, “Smart vest: Wearable multiparameter remote.
Physiological monitoring system,” Med. Eng. Phys., vol. 30, no. 4, pp. 466–477,
May 2008.
[8] Brian Dolan, "300K patients were remotely monitored in 2012" [Online: Accessed
24 /06/2017]:http://mobihealthnews.com/19963/report-about-300k-patients-were-
remotely-moni stored-in-2012.
[9] Laurie Orlov, " Ten Technologies from the 2013", mHealth Summit [Online:
Accessed 2206/2017] http://www.ageinplacetech.com/blog/ten-technologies-2013-
mhealth-summit.
[10] Real-time, "Can M2M solve the healthcare cost crisis", Online: [Accessed 24
/06/2017]: http://www.orange-business.com/en/magazine/technology/can-m2m-
solve-the-healthcare cost-crisis.
Page 7 - 15
[11] Rebecca Hill, public technology, " Large-scale cyber-attack hits hospitals across
England", [Accessed 29/06/2017]: https://www.publictechnology.net /articles/ne
ws/large-scale-cyber-attack-hits-hospitals-across-england.
[12] Akyildiz I.F., Jornet J.M., The Internet of Nano-Things, IEEE Wireless
Communication, vol. 17, no.6, 2010.
[13] 5G Vision: The 5G Infrastructure Public-Private Partnership: The Next Generation
of Communication Networks and Services, 5G-PPP, 2015; https://5g-ppp.eu/wp-
content/ uploads/2015/02/5G-Vision-Brochure-v1.pdf.
[14] Zhong Fan and S. Tan, "M2M communications for e-health: Standards, enabling
technologies, and research challenges," 2012 6th International Symposium on
Medical Information and Communication Technology (ISMICT), La Jolla, CA,
2012, pp. 1-4.
[15] 5G Vision: The 5G Infrastructure Public-Private Partnership: The Next Generation
of Communication Networks and Services, 5G-PPP, 2015; https://5g-ppp.eu/wp-
content/ uploads/2015/02/5G-Vision-Brochure-v1.pdf.
[16] C. Turcu and C. Turcu, "Internet of Things as a key enabler for sustainable
healthcare delivery," Procedia - Social and Behavioral Sciences, vol. 73, pp. 251-
256, 2013.
[17] Z. Fan, R. J. Haines, and P. Kulkarni, "M2M communications for E-health and smart
grid: an industry and standard perspective," in IEEE Wireless Communications, vol.
21, no. 1, pp. 62-69, February 2014.doi: 10.1109/MWC.2014.6757898.
[18] I. Kononenko, "Machine learning for medical diagnosis: history, state of the art and
perspective," Artificial Intelligence in Medicine, 23.1: 89-109, 2001.
[19] Relianthear, [ Accessed 29/06/2017]: http://reliantheart.com/ products /hear tassist
5-2/.
[20] W. Zhao, W. Chaowei, and Y. Nakahira, “ Medical application on the Internet of
Things,'' in Proc. IET Int. Conf. Commun. Technol. Appl. (ICCTA), Oct. 2011, pp.
660-665
[21] J. Swetina et al., “Toward a Standardized Common M2M Service Layer Platform:
Introduction to oneM2M,” IEEE Wireless Comm., vol. 21, no. 3, 2014, pp. 20–26.
[22] E. Kartsakli et al., “A Survey on M2M Systems for MHealth: A Wireless
Communications Perspective,” Sensors, vol. 14, no. 10, 2015, pp. 18009–18052.
Page 7 - 16
[23] J. Jarmakiewicz, K. Parobczak, and K. Maślanka, "On the Internet of Nano Things
in healthcare network," 2016 International Conference on Military Communications
and Information Systems (ICMCIS), Brussels, 2016, pp. 1-6.
[24] American Hospital Association: Hospitals. JAHA 35(15): 383–430 [Online:
Accessed16/07/2017]: http://www.cdc.gov/nchs/data/hus/2012/108.pdf.
[25] Gtxcorphttp, [ Accessed 29/06/2017]: http://gpssmartsole.com/gpssmartsole
[26] S. M. R. Islam, D. Kwak, M. H. Kabir, M. Hossain and K. S. Kwak, "The Internet
of Things for Health Care: A Comprehensive Survey," in IEEE Access, vol. 3, no,
pp. 678-708, 2015.
[27] M. Rayner, S. Allender and P. Scarborough, "Cardiovascular disease in Europe,"
Euro-pean Journal of Cardiovascular Prevention & Rehabilitation, vol. 16, no. 1, pp.
S43-S47, 2009.
[28] Avalere Health. Analysis of American Hospital Association Annual Survey data,
2011, for community hospitals. U.S. Census Bureau: National and State Population
Estimates, July 1, 2011. Links on page 2 & page 16/.
[29] MedSignals, [ Accessed 29/06/2017]: http://www.medsignals.com/.
[30] oneM2M Use Cases Collection, tech. report, OneM2M, 2013;
www.etsi.org/deliver/etsi_tr/118500_118599/118501/01.00.00_60/tr_118501v0
10000p.pdf.
[31] Petersen S, Rayner M, Wolstenholme J (2002) Coronary heart disease statistics:
heart failure supplement. London: British Heart Foundation.
Page 8 - 1
CHAPTER 8
CONCLUSION AND FUTURE WORK
Page 8 - 2
8.1 Conclusion
M2M and IoT technologies will make up a large part of the future of medical treatment.
Exploitation of the technology of implants and nano-network of sensors will enable remote
healthcare treatments and improve medical services. Wireless technology can increase the
flexibility of medical applications, besides enabling some applications such as remote
monitoring. Meanwhile, wireless technology could face challenges in the medical sector, such
as interference problems, battery consumption, and spectrum limitations. Cognitive radio as
smart technology can help to address and solve these challenges.
To meet the requirements of spectrum needs in CM2M networks a novel aggregation-based
spectrum assignment algorithm introduced. The algorithm maximizes spectrum utilization to
CM2M devices as a criterion to realize spectrum assignment. Moreover, the introduced
algorithm takes into account the realistic constraints of Co-Channel Interference and
Maximum Aggregation Span. Simulations validate the performance of the proposed algorithm,
and results are compared with algorithms available in the literature. The proposed algorithm
decreases the number of rejected devices and improves spectrum utilization of the CM2M
network. The developed algorithm increases the capacity of the network, which is vital for
CM2M networks.
Moreover, the work in this thesis also considers the energy efficiency in CM2M networks, to
prolong CM2M gateways battery life and increase the energy efficiency; Chapter 4 introduced
an energy efficient mechanism boost the energy efficiency in CM2M e-healthcare system by
35%. The proposed mechanism simultaneously considers the wait/switch trade-off regarding
channel switching probability and the sensing /throughput trade-off regarding the duration of
sensing time. The proposed mechanism addresses the collision between CM2MDs by
considering the transmission of devices to be zero when a collision occurs, and guarantees that
the given constraints regarding throughput and sensing reliability are always satisfied.
Furthermore, the algorithm considers when no channel available for transmission CM2M
devices will sleep for a period of time-saving more energy and will go back to
active/transmission mode when more data need to be transmitted.
Furthermore, the thesis in Chapter 5 and 6 considered antenna selection sensing scheme and
learning algorithms for more sensing accuracy and better channel selection scheme for CM2M
network environment. The developed scheme and algorithm improved the mechanism give in
Page 8 - 3
Chapter 4 and increased the energy efficiency of CM2M network by 45%. Furthermore, the
mechanism boosts the accuracy of CM2M devices, which leads to more probability of
detection and less sensing time required from CM2M to sense the available channels. Chapter
7 explored the future of M2M technology in medical health, real-world examples given and
explained. Besides, the Chapter explored the security aspects that could affect the service of
M2M communication in medical health applications and finally recommendations given to
cope with the security vulnerabilities.
8.2 Future work
1- For future work, from the CM2M spectrum efficiency perspective and based on the results
of this thesis the impact of the various parameters could be explored by using a genetic
algorithm to solve the introduced utilisation function; population size, crossover rate and
mutation rate; in addition, future work could consider developing a genetic algorithm
based method to assign spectrum to CM2M devices in an energy efficient manner [1].
2- For future work, from a CM2M energy efficiency perspective, different sensing and
accessing techniques could be tested and designed to improve sensing accuracy and reduce
interference problems. Match filter or Feature detections schemes can be used to ensure
the low cost of the design due to M2M communications requirements of low cost and
simplicity [2, 3]. Furthermore, additional monitoring techniques could be added to the
current models such as Receiver Statistics and Energy Ratio to improve channel
observation and increase the sensing and accessing accuracy due to selecting better
channels [4, 5].
3- Exploiting CM2M technology in wireless healthcare applications, as proved in this thesis,
cognitive radio as smart technology can boost spectrum efficiency and decrease energy
consumption. More studies on the use of cognitive radio in M2M healthcare applications
could help to solve many problems of wireless M2M communications in healthcare
systems.
4- Low latency is one of the important requirements of healthcare applications. For instance,
in telesurgery services, latency in wireless communications has a significant effect on the
operation and the process of robotic instruments. Low latency (less than 200 ms) is
considered good and amenable for the next generation of telesurgery systems. Cognitive
Page 8 - 4
radio as smart technology can be designed to reduce the latency in operating systems (e.g.,
telesurgery systems) that will lead to unleashing further mobile applications working with
a fixed latency rate [6]. Reducing latency using smart technologies such as cognitive radio
will enable urgent and specialist operations to be carried out remotely by specialist
surgeons anywhere in the world.
5- Due to the huge number of biomedical sensors with M2M capabilities, machine-generated
data could soon exceed network capacity. Thus, efficient schemes and techniques should
be developed to adapt to such challenges. Based on the literature review of this thesis,
cognitive radio as intelligent technology can be adapted to work with a huge number of
devices [7]. The SDR capability in CR enables it to manage a huge number of SUs with
good communication quality and reliability. Weighted cooperative spectrum sensing
schemes could be used to increase the optimality of secondary user (M2M devices) with a
guarantee of good throughput and reliability constraints.
Page 8 - 5
References for Chapter 8
[1] H. Shariatmadari et al., “Machine-Type Communications: Current Status and Future
Perspectives toward 5G Systems,” IEEE Comm., vol. 53, no. 9, 2015, pp. 10–17.
[2] L. Ma, Y. Li, and A. Demir, “Matched filtering assisted energy detection for sensing
weak primary user signals,” in Proc. IEEE Int. Conf. Acoust. Speech Signal Process.,
Kyoto, Japan, Mar. 2012, pp. 3149–3152.
[3] A. Nasser, A. Mansour, K. C. Yao, H. Charara, and M. Chaitou, “Efficient spectrum
sensing approaches based on waveform detection,” in Proc. Int. Conf. e-Technol.
Netw. Develop., Beirut, Lebanon, Apr. 2014, pp. 13–17.
[4] S. W. Boyd, J. M. Frye, M. B. Pursley, and T. C. Royster, IV, “Spectrum monitoring
during a reception in dynamic spectrum access cognitive radio networks,” IEEE
Trans. Commun., vol. 60, no. 2, pp. 547–558, Feb. 2012.
[5] A. Ali and W. Hamouda, “Spectrum monitoring using energy ratio algorithm for
OFDM-based cognitive radio networks,” IEEE Trans. Wireless Commun., vol. 14, no.
4, pp. 2257–2268, Apr. 2015.
[6] Y. C. Chen, I. W. Lai, K. C. Chen, W. T. Chen and C. H. Lee, "Transmission latency
and reliability trade-off in path-time coded cognitive radio ad hoc networks," 2014
IEEE Global Communications Conference, Austin, TX, 2014, pp. 1084-1089. doi:
10.1109/GLOCOM.2014.7036953.
[7] T. Chakraborty, I. S. Misra and T. Manna, "Design and Implementation of VoIP Based
Two-Tier Cognitive Radio Network for Improved Spectrum Utilization," in IEEE
Systems Journal, vol. 10, no. 1, pp. 370-381, March 2016.
APPENDIX A
MEDICAL TELEMETRY FREQUENCY
BAND
Medical telemetry frequency Band Description
Inductive coupling devices <1MHz
Wireless Medical Telemetry System 608-614 MHz, 1395-1400 MHz, 1427-
1429.5 MHz
Medical device radio communication
service 401 to 406 MHz
802.11a WIFI 5 GHz
802.11b WIFI 2.4 GHz
802.11g WIFI 2.4 GHz
802.11n WIFI 2.4 / 5 GHz
802.15.1 Bluetooth Class-I 2.4 GHz
802.15.1 Bluetooth Class-II 2.4 GHz
802.15.4 (Zigbee) 868 MHz, 915 MHz, 2.4 GHz
APPENDIX B
MATLAB CODES FOR A NUMBER OF
SIMULATIONS
%This is the main code of the Genetic Algorithm (GA) based in Spectrum
%aggregation (SA) configured for Maximum Satisfaction Algorithm and Random
%Channel Assignment Algorithm (RCA) (Chapter 3)
% A variable declared as global in multiple functions (and optionally the base %workspace) share a single copy of that variable. Any assignment you
make to %that variable in one function is available to all functions that
declare it %global.
%INPUTS
% Define variables
% PU_Space = primary user space
& ra_cr = probability of Collison
% MAS = Maximum Aggregation Span
% Len_Chro = Length of the chromosome
% User_Req_Bw = User request bandwidth
% gen_num = generation number
% relative_white_space = relative white space to each channel
% C = matrix C
clear
clc
global ra_cr PU_space C white_space N M L MAS M_new N_Non_Inter M_Non_Inter
Len_Chro mutationRate size_pop Len_Chro L_new C_new User_Req_Bw
relative_white_space gen_num
%<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<
iteration_num=20; % Iteration number
N=10; %Number of CM2M device
M=30; %Number of Overleaping channels
MAS=40;
size_pop=30; % population size
gen_num=50; % Number of Generation
mutationRate=.01;
% how_man=1;
% ra_cr=.6;
%>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>
Max_User_Req=20;
Max_white_space=10;
Max_PU_space=20;
white_space=round((Max_white_space-1)*rand(1,M)+1);
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
ashg=0;
for N=5:5:60
ashg=ashg+1;
Tx=0;
Rx=0;
Hx=0;
Zx=0;
n_LL=0;
BBX=0;
CCX=0;
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
for kkee=1:iteration_num % loop for a number of iteration circle
L=ones(N,M);
C=zeros(N,N,M);
% rng default
% white_space=round((Max_white_space-1)*rand(1,M)+1);
User_Req_Bw=round((Max_User_Req-1)*rand(1,N)+1);
PU_space=round((Max_PU_space-1)*rand(1,M-1)+1);
M_new=sum(white_space);
L_new=[];
C_new=[];
%..........................................................................
for n1=1:N
for n2=1:N
for n3=1:M
if n1==n2
C(n1,n2,n3)=0;
else
C(n1,n2,n3)=1;
end
end
end
end
totall_bw_msa=0;
num_of_rej_msa=0;
[totall_bw_msa num_of_rej_msa]=MSA;
BBX=BBX+totall_bw_msa;
CCX=CCX+num_of_rej_msa;
TTotal_of_given_bw_random=0;TTotal_of_given_bw_aggreg=0;
NNumber_of_rejected_users_random=0;
NNumber_of_rejected_users_aggreg=0;
%rng default
[TTotal_of_given_bw_random NNumber_of_rejected_users_random
assign_result_of_random] = Rand_chan_ass;
%rng default
[TTotal_of_given_bw_aggreg NNumber_of_rejected_users_aggreg
assign_result_of_aggreg ] = aggre_aware;
Tx=Tx+TTotal_of_given_bw_random;
Rx=Rx+TTotal_of_given_bw_aggreg;
Zx=Zx+NNumber_of_rejected_users_random;
Hx=Hx+NNumber_of_rejected_users_aggreg;
n_LL=sum(User_Req_Bw)/sum(white_space)+n_LL;
end
%-------------------------------------------------------------------
--
Total_of_given_bw_random(ashg)=Tx/iteration_num;
Total_of_given_bw_aggreg(ashg)=Rx/iteration_num;
%------------------------------------------------------------------
---
Number_of_rejected_users_random(ashg)=Zx/iteration_num;
%
Number_of_rejected_users_aggreg(ashg)=Hx/iteration_num;
TTtotall_bw_msa(ashg)=BBX/iteration_num;
NNnum_of_rej_msa(ashg)=CCX/iteration_num;
n_L(ashg)=n_LL/iteration_num;
end
%Plotting figures
figure(1)
hold on
zzxx=sum(white_space)*ones(1,12)
NNMM=[5:5:60];
ccrrt=NNMM-Number_of_rejected_users_random;
plot(n_L,zzxx,'-
rs','MarkerEdgeColor','yellow','MarkerFaceColor','r','MarkerSize',5)
plot(n_L,Total_of_given_bw_aggreg,'-
rs','MarkerEdgeColor','r','MarkerFaceColor','r','MarkerSize',5)
plot(n_L,TTtotall_bw_msa,'-
ks','MarkerEdgeColor','green','MarkerFaceColor','k','MarkerSize',5)
plot(n_L,Total_of_given_bw_random,'-
ks','MarkerEdgeColor','k','MarkerFaceColor','k','MarkerSize',5)
figure(2)
hold on
plot(n_L,(Number_of_rejected_users_random),'-
ks','MarkerEdgeColor','k','MarkerFaceColor','k','MarkerSize',5)
plot(n_L,NNnum_of_rej_msa,'-
ks','MarkerEdgeColor','green','MarkerFaceColor','k','MarkerSize',5)
plot(n_L,Number_of_rejected_users_aggreg,'-
rs','MarkerEdgeColor','r','MarkerFaceColor','r','MarkerSize',5)
% This code for the main creation of different population size with Maximum
%Aggregation Span algorithm. During the initialization process, the initial
%population is randomly generated based on a binary coding mechanism.
%(Chapter 3)
% New functions may be added to MATLAB's vocabulary if they
% are expressed in terms of other existing functions. The
% commands and functions that comprise the new function must
% be put in a file whose name defines the name of the new
% function, with a filename extension of '.m'. At the top of
% the file must be a line that contains the syntax definition
% for the new function. For example, the existence of a file
% on disk called stat.m with:
%
% function [mean,stdev] = stat(x)
% %STAT Interesting statistics.
% n = length(x);
% mean = sum(x) / n;
% stdev = sqrt(sum((x - mean).^2)/n);
% Define variables
% PU_Space = primary user space
& ra_cr = probability of Collison
% MAS = Maximum Aggregation Span
% Len_Chro = Length of the chromosome
% User_Req_Bw = User request bandwidth
% gen_num = generation number
% relative_white_space = relative white space to each channel
% C_New = updated matrix C
% N= Number of CM2M devices
% M= Number of overlapping channels
function Population = main_creation( Len_Chro, MSR )
global N M_new M_Non_Inter size_pop N_Non_Inter User_Req_Bw L_new
relative_white_space MAS C_new white_space
Population=[];
%--------------------------------------------------------------------------
for n1=1:(size_pop)
assignment_matrix=zeros(N,M_new);
p=[];p=randperm(N);
%========================================================================
for n9=1:N
fff=zeros(1,M_new);
ttt=zeros(1,M_new);
rrr=zeros(1,M_new);
yy=zeros(1,M_new);
kk=User_Req_Bw(p(n9));
xx=zeros(1,M_new);
ll=[]; ll= find(L_new(p(n9),:));
xx(1,ll)=1;
hhhh=[];ttt=[];
======================================================================
for n15=1:(n9-1)
rrr(1,:)=C_new(p(n9),p(n15),:);
yy=assignment_matrix(p(n15),:);
fff=rrr.*xx.*yy;
hhhh=xx-fff;
xx=hhhh;
ttt=find(hhhh);
ll=[];
ll=ttt;
end
%========================================================================
hh=size(ll);
if hh(2)>=kk
ppp=[];
www=[];
ppp=sort(ll);
www=relative_white_space(ppp);
var_temp3=0;
iiii=[];
%========================================================================
if kk~=1
for n10=1:(hh(2)-1)
var_temp1=1;
for n11=(n10+1):hh(2)
if (www(n11)-www(n10))<=MAS
var_temp1=var_temp1+1;
if var_temp1==kk
var_temp3=var_temp3+1;
%in this if we can find without any aggregation
iiii(1,var_temp3)=ppp(n10);
iiii(2,var_temp3)=ppp(n11);
iiii(3,var_temp3)=n11-n10;
iiii(4,var_temp3)=n10;
end
if var_temp1>kk
iiii(1,var_temp3)=ppp(n10);
iiii(2,var_temp3)=ppp(n11);
iiii(3,var_temp3)=n11-
n10;iiii(4,var_temp3)=n10;
end
end
end
end
end
%========================================================================
if kk==1
for n10=1:hh(2)
var_temp3=var_temp3+1;
iiii(1,var_temp3)=ppp(n10);iiii(2,var_temp3)=ppp(n10);
iiii(3,var_temp3)=0;
iiii(4,var_temp3)=n10;
end
end
%========================================================================
mm=size(iiii);
if mm(2)~=0
uuuu=[];
hhhhh=[];
%add one for number of space minus one for decreasing start point
hhhhh=iiii(2,:)-iiii(1,:)+1-1;
for n13=1:mm(2)
uuuu(n13)=nchoosek(hhhhh(n13),(kk-1));
end
uuuu2=sum(uuuu);
uuuu3=uuuu/uuuu2;
jjjjj=[];
for n14=1:mm(2)
jjjjj(n14)=sum(uuuu3(1,1:n14));
end
rrrr=rand(1);
jjjjj=[0,jjjjj];
yyy=[];
yyy=find(jjjjj>rrrr);
cho_section=yyy(1)-1;
pp=[];
%pp=randperm(iiii(2,cho_section)-iiii(1,cho_section));
pp=randperm(iiii(3,cho_section));
ww=[];
% ww=pp+iiii(1,cho_section);
ww=pp+iiii(4,cho_section);
rr=[];
ee=[];
%rr=ww(1:(kk-1));
rr=ppp(ww(1:(kk-1)));
rr=[iiii(1,cho_section),rr];
ee=sort(rr);
assignment_matrix(p(n9),ee)=1;
end
end
end
for n4=1:Len_Chro
chro(n4)=assignment_matrix(N_Non_Inter(n4) ,M_Non_Inter(n4));
end
Population=[chro;Population];
end
end
%This code to for Maximum Satisfaction Algorithm (MSA) to show the total
%bandwidth of MSA and the number of Rejected devices in MSA. The proposed
%algorithm takes into consideration interference to the PUs, CCI among SUs,
%and MAS to aggregate whitespaces (Chapter 3)
function [totall_bw_msa num_of_rej_msa ] = MSA
global M MAS Max_white_space PU_space User_Req_Bw
global PU_space white_space N M L MAS M_new N_Non_Inter M_Non_Inter Len_Chro
mutationRate size_pop Len_Chro L_new C_new User_Req_Bw relative_white_space
gen_num
%.....................................................................
%.......................................................................
% Define variables
% totall_bw_msa = total bandwidth of MSA
% num_of_rej_msa = number of rejected devices
% PU_Space = primary user space
& ra_cr = probability of Collison
% MAS = Maximum Aggregation Span algorithm
% Len_Chro = Length of the chromosome
% User_Req_Bw = User request bandwidth
% gen_num = generation number
% relative_white_space = relative white space to each channel
% C_New = updated matrix CCI
% N= Number of CM2M devices
% M= Number of overlapping channels
accepted_users=zeros(4,N);
assignment_matrix=zeros(N,M);
accepted_users=zeros(4,N);
assignment_matrix=zeros(N,M);
V=0;modi_PU_space=[PU_space,2*MAS];
SC=[];
for n1=1:M
if modi_PU_space(n1)>=MAS
V=V+1;
%V is total number of CS(areas that separated because of PU space bigger
than mas)
SC(V)=n1;
%each element the number of PU space that finish each CS
end
end
[sa,so]=size(SC);
%.......................................................................
CS=[];
for n1=2:so
tem_r=0;
xxx=SC(n1-1)+1;
%the number of first white space after previous pu space divider
yur=1;
for hhtt=(SC(n1-1)+1):SC(n1)
bnrq(n1,yur)=xxx;
% rows show number of CS and coulmuns non zero are white space num in that
rwo
tem_r=white_space(xxx)+tem_r;
CS(n1)=tem_r;
% it is same size of sc but each column show how much white space is
available in that column number SC
xxx=xxx+1;
yur=yur+1;
end
end
tem_r=0;
xxx=1;yur=1;
for hhtt=1:SC(1)
bnrq(1,yur)=xxx;
tem_r =white_space(xxx)+tem_r;
CS(1)=tem_r;
xxx=xxx+1;
yur=yur+1;
end
%.......................................................................
R_CS=[];
R_CS=[CS;1:V];%make label for CS areas
R_User_Req_Bw=[];R_Des_User_Req_Bw=[];
R_User_Req_Bw=[User_Req_Bw;1:N];%make label for user req
d1=[];d2=[];
[d1,d2] = sort(R_User_Req_Bw(1,:),'descend');
R_Des_User_Req_Bw=R_User_Req_Bw(:,d2);
% Des_User_Req_Bw=sort(User_Req_Bw,'descend');
d1=[];d2=[];R_Asc_CS=[];
[d1,d2] = sort(R_CS(1,:));
R_Asc_CS=R_CS(:,d2);
% Asc_CS=sort(CS);
dytu=sort(white_space);
%.......................................................................
new_primary=PU_space;
new_white_space=[white_space;1:(M)];
M_nnn=M;
fla_empty_pu=0;
for jj=1:N
M_nnn
new_primary
flag_assinged=0;
for r=1:V
if (flag_assinged~=1)&&(fla_empty_pu~=1)
x_d=R_Asc_CS(2,r);
tem_er=[];
tem_ewer=[];
tem198=[];
tem_ewee33=[];
tem012=[];
t1=[];t2=[];PPP=[];
%..............................................................
tem_er=find(bnrq(x_d,:)>0);
[sa3,so3]=size(tem_er);
lljj=bnrq(x_d,tem_er);%so3 is always non zero
zse=[];
for ert=1:so3
zse(ert)=find(new_white_space(2,:)==lljj(ert));
%because new white space changes to find out
end
t1=new_white_space(1,zse);
t2=new_white_space(2,zse);%in fact t2 and zse are same
% t1=new_white_space(1,bnrq(x_d,tem_er));
% t2=new_white_space(2,bnrq(x_d,tem_er));
tem012= t1-R_Des_User_Req_Bw(1,jj);
tem_ewer= find(tem012==0);
tem198=t2(1,tem_ewer);
[sa,so]=size(tem198);
%..............................................................
if so~=0
assignment_matrix(R_Des_User_Req_Bw(2,jj),t2(tem_ewer(1)))=1;
flag_assinged=1;
% trei=new_white_space(1,tem_ewer(1));
trei=R_Des_User_Req_Bw(1,jj);
tnew_primary=[];
tnew_primary=new_primary;
new_primary=[];
[sa2,so2]=size(tnew_primary);
accepted_users(2,jj)=t2(tem_ewer(1)) ;
accepted_users(3,jj)=t2(tem_ewer(1)) ;
accepted_users(4,jj)=R_Des_User_Req_Bw(1,jj);
%__________________________________________________________________________
_
if so2>=3
% if
(tem_ewer(1)>2)&&(tem_ewer(1)<so2+1-2)
if (zse(tem_ewer(1))>2)&&(zse(tem_ewer(1))<=so2+1-2)
new_primary=[tnew_primary(1:zse(tem_ewer(1))-
2),tnew_primary(1,zse(tem_ewer(1))-
1)+trei+tnew_primary(1,zse(tem_ewer(1))),tnew_primary((zse(tem_ewer(1))+1):
end)];
new_white_space(:,zse(tem_ewer(1)))=[];
% elseif
zse(tem_ewer(1))==3
%
% new_white_space(:,3)=[];
%
elseif zse(tem_ewer(1))==2
new_primary=[tnew_primary(1,1)+trei+tnew_primary(1,2),tnew_primary(3:end)];
new_white_space(:,2)=[];
elseif zse(tem_ewer(1))==1
tnew_primary(:,1)=[];new_primary=tnew_primary;
new_white_space(:,1)=[];
elseif zse(tem_ewer(1))==(so2+1)
tnew_primary(:,so2)=[];new_primary=tnew_primary;
new_white_space(:,so2+1)=[];
elseif zse(tem_ewer(1))==(so2)
new_primary=[tnew_primary(1,1:so2-
2),tnew_primary(1,so2-1)+trei+tnew_primary(1,so2)];
new_white_space(:,so2)=[];
elseif (so2==4)&&(zse(tem_ewer(1))==3)
new_white_space(:,3)=[];
new_primary=[tnew_primary(1,1),tnew_primary(1,2)+trei+tnew_primary(1,3),tne
w_primary(1,4)];
end
end
%_________________________________________________________________________
if so2==2
if zse(tem_ewer(1))==2
new_primary=[tnew_primary(1,1)+trei+tnew_primary(1,2)];
new_white_space(:,2)=[];
elseif zse(tem_ewer(1))==1
new_primary=[tnew_primary(1,2)];
new_white_space(:,1)=[];
elseif zse(tem_ewer(1))==3
new_primary=[tnew_primary(1,1)];
new_white_space(:,3)=[];
end
end
%_________________________________________________________________________
if so2==1
if zse(tem_ewer(1))==2
%
new_primary=[tnew_primary(1,2)];
new_primary=[];
new_white_space(:,2)=[];
elseif zse(tem_ewer(1))==1
%
new_primary=[tnew_primary(1,1)];
new_primary=[];
new_white_space(:,1)=[];
end
fla_empty_pu=1;
end
%_________________________________________________________________________
% v ro ham tagir kon
%az loop for kharaj sho
M_nnn=M_nnn-1;
else
tem_ewee33=find(tem012<0);
tem012(1,tem_ewee33)=789087;
frty=min(tem012);
if frty~=789087
PPP=find(tem012==frty);
assignment_matrix(R_Des_User_Req_Bw(2,jj),t2(PPP(1)))=1;
accepted_users(2,jj)=t2(PPP(1)) ;
accepted_users(3,jj)=0 ;
accepted_users(4,jj)=R_Des_User_Req_Bw(1,jj);
flag_assinged=1;
tnew_primary=[];
tnew_primary=new_primary;
new_primary=[];
[sa2,so2]=size(tnew_primary);
%__________________________________________________________________________
_
if zse(PPP(1))==1
new_primary=tnew_primary;
new_white_space(1,1)=new_white_space(1,1)-
R_Des_User_Req_Bw(1,jj);
%M_nnn=M_nnn-1;
elseif zse(PPP(1))==(so2+1)
new_primary=[tnew_primary(1:zse(PPP(1))-
2),tnew_primary(1,zse(PPP(1))-1)+R_Des_User_Req_Bw(jj)];
new_white_space(1,zse(PPP(1)))=new_white_space(1,zse(PPP(1)))-
R_Des_User_Req_Bw(1,jj);
else
new_primary=[tnew_primary(1:zse(PPP(1))-
2),tnew_primary(1,zse(PPP(1))-
1)+R_Des_User_Req_Bw(jj),tnew_primary(zse(PPP(1)):end)];
% v ro ham tagir kon
new_white_space(1,zse(PPP(1)))=new_white_space(1,zse(PPP(1)))-
R_Des_User_Req_Bw(1,jj);
%az loop for kharaj sho
% M_nnn=M_nnn-1;
end
end
end
if flag_assinged==1
accepted_users(1,jj)=1 ;
end
end
en
%.......................................................................
if fla_empty_pu~=1
%*****************
modi_PU_space=[];
modi_PU_space=[new_primary(1,:),2*MAS];
%.......................................................................
V=0;
SC=[];
for n1=1:M_nnn
if modi_PU_space(n1)>=MAS
V=V+1;
SC(V)=n1;
end
end
[sa,so]=size(SC);
%.......................................................................
CS=[];bnrq=[];
for n1=2:so
tem_r=0;
xxx=SC(n1-1)+1;
yur=1;
for hhtt=(SC(n1-1)+1):SC(n1)
% bnrq(n1,yur)=xxx;
bnrq(n1,yur)=new_white_space(2,xxx);
tem_r=new_white_space(1,xxx)+tem_r;
CS(n1)=tem_r;
xxx=xxx+1;
yur=yur+1;
end
end
%.......................................................................
tem_r=0;
xxx=1;yur=1;
for hhtt=1:SC(1)
% bnrq(1,yur)=xxx;
bnrq(1,yur)=new_white_space(2,xxx);
tem_r=new_white_space(1,xxx)+tem_r;
CS(1)=tem_r;
xxx=xxx+1;
yur=yur+1;
end
%.......................................................................
R_CS=[];
R_CS=[CS;1:V]; d1=[];d2=[];
[d1,d2] = sort(R_CS(1,:));
R_Asc_CS=R_CS(:,d2);
%*****************
else
tttttt=111111111111111111111
for kkiioo=jj+1:N
if R_Des_User_Req_Bw(1,kkiioo)<new_white_space(1,1)
assignment_matrix(R_Des_User_Req_Bw(2,kkiioo),new_white_space(2,1))=1;
accepted_users(1,kkiioo)=1 ;
accepted_users(2,kkiioo)=666 ;
accepted_users(3,kkiioo)=0;
accepted_users(4,kkiioo)=R_Des_User_Req_Bw(1,kkiioo);
flag_assinged=1;
new_white_space(1,1)=new_white_space(1,1)-
R_Des_User_Req_Bw(1,kkiioo);
elseif R_Des_User_Req_Bw(1,kkiioo)==new_white_space(1,1)
assignment_matrix(R_Des_User_Req_Bw(2,kkiioo),new_white_space(2,1))=1;
accepted_users(1,kkiioo)=1 ;
accepted_users(2,kkiioo)=666 ;
accepted_users(3,kkiioo)=666;
accepted_users(4,kkiioo)=R_Des_User_Req_Bw(1,kkiioo);
flag_assinged=1; fla_empty_pu=1;
new_white_space(1,1)=new_white_space(1,1)-
R_Des_User_Req_Bw(1,kkiioo);
end
end
end
end
%%%%%%TEST
for uiyt=1:M
tture1=[];
tture1=find(accepted_users(2,:)==uiyt);
[sa,so]=size(tture1);
if so==0
rqqp(1,uiyt)=99999;
else
iiop=sum(white_space(uiyt));
zzqq=sum(R_Des_User_Req_Bw(1,tture1));
rqqp(1,uiyt)=zzqq-iiop;
end
end
totall_bw_msa=sum(R_Des_User_Req_Bw(1,find(accepted_users(1,:)==1)));
num_of_rej_msa=N-sum(accepted_users(1,:));
%NL(ilo)=sum(User_Req_Bw)/sum(white_space);
%%%%%%TEST
end
%This code for Maximizing Sum of Reward algorithm in cognitive radio systems
%( Chapter 3) To achieve higher spectrum efficiency and faster convergence, %after each generation the MSRA whenever possible randomly assigns all
%unassigned spectrum to remaining SUs,
% Define variables
% Chro = Chromosome of GA
% totall_bw_msa = total bandwidth of MSA
% num_of_rej_msa = number of rejected devices
% PU_Space = primary user space
& ra_cr = probability of Collison
% MAS = Maximum Award summation algorithm
% Len_Chro = Length of the chromosome
% User_Req_Bw = User request bandwidth
% gen_num = generation number
% relative_white_space = relative white space to each channel
% C_New = updated matrix C
% N= Number of CM2M devices
% M= Number of overlapping channels
function MSR_OUT = MSR(Chro)
global N M_new N_Non_Inter M_Non_Inter Len_Chro User_Req_Bw
A=zeros(N,M_new);
for n4=1:Len_Chro
A(N_Non_Inter(n4),M_Non_Inter(n4))=Chro(n4);
end
tem_var1=0;
tem_var=0;
for n1=1:N
tt=sum(A(n1,:));
if tt==0
tem_var=tem_var+1;
else
tem_var1=tem_var1+User_Req_Bw(n1);
end
end
%max bw
MSR_OUT=-tem_var1;
%min rejection
%MSR_OUT=tem_var;
end
%This code for Monte Carlo simulation detection, when the primary signal is
%real Gaussian signal and noise is % additive white real Gaussian (Chapters
%4, 5, 6 )
clc
close all
clear all
L = 1000;
snr_dB = -10; % SNR in decibels
snr = 10.^(snr_dB./10); % Linear Value of SNR
Pf = 0.01:0.01:1; % Pf = False Alarm Probability
for m = 1:length(Pf)
m
i = 0;
for kk=1:10000 % Monte Carlo Simulations Numbers
n = randn(1,L); %AWGN noise with mean 0 and variance 1
s = sqrt(snr).*randn(1,L); % Real valued Gaussina Primary User Signal
y = s + n; % SU Received signal
energy = abs(y).^2; % Energy of received signal over Ɲ samples
energy_fin =(1/L).*sum(energy); % Test Statistic for the energy detection
thresh(m) = (qfuncinv(Pf(m))./sqrt(L))+ 1; if(energy_fin >= thresh(m))
% Check whether the received energy is greater than threshold, if so,
increment Pd
(Probability of detection)
counter by 1
i = i+1;
end
end
Pd(m) = i/kk;
end
plot(Pf, Pd)
hold on
%This Code for Spectrum Sensing for CM2M gateways to sense the available
%channels with the consideration to the primary users using QPSK Modulation
%and BPSK Modulation (Chapter 4, 5, ,6 ).
clear all;
%
% PARAMETERS
%
freq = 200; %operating frequency
Fs = 20*f; %sampling frequency
L=100; % Number of samples per symbol period
Ts = 1/Fs; % Sampling period
T = Ts:Ts:1/f;
alpha=0.5; % Roll-off factor for the (square-root) raised cosine filters
N=8*L; % N+1 is the length of the square-root raised-cosine filter.
sigma_v=0; % Standard deviation of channel noise
h=1; % Channel impulse response
%
%SOURCE: Take input data from user for transmission
%
pt_dt = input('Data you want to send:','s');
R = isempty(pt_dt);
if R == 1
pt_dt = 'Waleed Ejaz';
else
pt_dt = pt_dt;
end
display(pt_dt);
RR = double(pt_dt);
bb = 1;
Rp = dec2bin(RR,7);
[TA TC] = size(Rp);
for ll = 1:1:TA
for lg = 1:1:TC
msg(bb) = Rp(ll,lg);
bb = bb + 1;
end
end
rt = 1; ht = 1;
for ls = 1:1:TA
for ll = 1:2:(TC-1)
Inp_msg(rt,(ht:ht+1)) = Rp(ls,(ll:ll+1));
rt = rt + 1;
end
end
%
% Transmit Filter
%
pT=f_sr_cos_p(N,L,alpha); % Transmit filter:
xT=conv(f_expander(msg,L),pT); % Transmit signal
%
% Modulation
%
display('Select Type of Modulation');
display('1. BPSK');
display('2. QPSK');
Mod_Type = input('Plz Enter the Type of Modulation :','s');
Carrier = [];
%
% BPSK Modulation
%
if (Mod_Type=='1')
display('Binary PSK');
for ii = 1:1:length(T)
car1(ii) = sin((2*pi*freq*T(ii))); %CARRIER TO BE TRANSMITTED
end
for ii = 1:1:length(xT)
if xT(ii) == '0'
car = -1*car1;
else
car = 1*car1;
end
Carrier = [Carrier car];
end
%
% QPSK Modulation
%
else if(Mod_Type=='2')
for ii = 1:1:length(T)
car1(ii) = sin((2*pi*freq*T(ii))+360); %CARRIER TO BE TRANSMITTED
car2(ii) = sin((2*pi*freq*T(ii))+90); %CARRIER TO BE TRANSMITTED
car3(ii) = sin((2*pi*freq*T(ii))+180); %CARRIER TO BE TRANSMITTED
car4(ii) = sin((2*pi*freq*T(ii))+270); %CARRIER TO BE TRANSMITTED
end
for ii = 1:1:length(Inp_msg)
if Inp_msg(ii) == '00'
car = car1;
else if Inp_msg(ii) == '01'
car = car2;
else if Inp_msg(ii) == '10'
car = car3;
else if Inp_msg(ii) == '11'
car = car4;
end
end
end
end
Carrier = [Carrier car];
end
if true
% code
end end % end of if
end %end of else if
%
if true
% code
end
% CHANNEL
%
xR=conv(h,Carrier);
if true
% code
end
xR=xR+sigma_v*randn(size(xR)); % Received signal
%This code for Spectrum Sensing and sharing Algorithm. This algorithm to
%sense the available channels and switch/sleep off when the sensed channels
%busy (Chapters 5)..more details please read Chapter 5.
% System Parameters/Inputs.
fs=46000; %Sample frequency Hz
X=0.1; %the probability of false alarm
SNR=10; %single to Nosie ratio
Bo=650; %bandwidth
ps=1; %probability of channel been busy
M=6; %number of available channels
S=8; %time
T=0.8; %frame time
Es=0.04; %Energy of Switching
Et=0.0695; %Energy of Transitions
y=10; %SNR of Primary user
p=0.40; %place holder
Pdt=0.9; %probability of detection
Pft=0.1; %Target probability of false alarm
% Sensing and Sharing Algorithm
Co=log2(1+SNR);
R0=(1-p)*Co;
R=U*R0*2
D=X*S;
Mue=D/D+S;
OOO=qfuncinv(Pft);
SSS=qfuncinv(Pdt);
pc1=(1-p)*(1-Pft);
pc2=p*(1-Pdt);
Pb=((Pdt*p+Pft*(1-p)))^(M-1);
P1=(1-p)*(1-0.1)+p*(1-0.9);
P3=(1-P1)*(1-Pb)*(1-ps)*(1-(0.5*(1-ps))); %Probability of Switching
PT=(P1+P3);
N=S/(PT*T);
t2=T-Bo*(1-exp(-(T/Bo)));
Bt=(1-(t2/T))*Co*T;
B1=pc1*Bt;
B2=P3*(1-Pe)*Bt;
tsoptr=(B1+B2-T*R)/R;
PWT=(1-P1)*Pb+(1-P1)*(1-Pb)*ps;
tsoptd=(D-N*T*PWT)/N;
Pt=P1+P3;
TS=S/Pt;
tsMIN=(1/fs*(y)^2)*(OOO-SSS*sqrt(2*y+1))^(2); % min sensing time
Power=(N*tsMIN*Es+N*P3*Jsw+S*Et)*2000; % Total power consumption
avgR=((B1+B2)/(tsMIN+T))*2;
% Ploting the results
psoptmal=1-(tsMIN+T)*R-pc1*Bt/Bt*(1-P1)*(1-Pe)*(1-Pb);
gg=(Mue*T-(1-Mue)*tsMIN)/(T*(1-P1)*(1-Pb))-(Pb/1-Pb);
avgD=N*tsMIN+N*T*PWT;
hold on
plot(U,Power,'--s')